# Installing the libraries with the specified version
!pip install --no-deps tensorflow==2.18.0 scikit-learn==1.3.2 matplotlib===3.8.3 seaborn==0.13.2 numpy==1.26.4 pandas==2.2.2 -q --user --no-warn-script-location

     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 61.0/61.0 kB 1.4 MB/s eta 0:00:00
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 615.5/615.5 MB 852.3 kB/s eta 0:00:00
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 10.8/10.8 MB 89.7 MB/s eta 0:00:00
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 11.6/11.6 MB 83.3 MB/s eta 0:00:00
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 18.0/18.0 MB 70.4 MB/s eta 0:00:00

# Import libraries for data manipulation
import numpy as np
import pandas as pd

# Import libraries for data visualization
import matplotlib.pyplot as plt
import seaborn as sns
import plotly.express as px

# Set default Seaborn theme for plots
sns.set()

# Suppress warnings to prevent them from being displayed during code execution
import warnings
warnings.filterwarnings('ignore')

# Enable the inline plotting of matplotlib figures directly within the notebook
%matplotlib inline

# Set the display format for float values to three decimal places
pd.set_option('display.float_format', lambda x: '%.4f' % x)

# Removes the limit for the number of displayed columns
pd.set_option("display.max_columns", None)
# Sets the limit for the number of displayed rows
pd.set_option("display.max_rows", 100)

# Distribution Libraries for statistical tests
import scipy.stats as stats
from scipy.stats import norm, uniform, binom, expon, t
from scipy.stats import ttest_1samp, ttest_ind

# Libraries for model building and evaluation
import statsmodels.api as sm
from sklearn.model_selection import train_test_split, GridSearchCV, RandomizedSearchCV, StratifiedKFold, cross_val_score

# Libraries to get different metric scores
from sklearn import metrics
from sklearn.metrics import (
    confusion_matrix,
    accuracy_score,
    precision_score,
    recall_score,
    f1_score,
    roc_auc_score,
    ConfusionMatrixDisplay,
    classification_report
)

# Libraries for model building (Logistic Regression, Decision Tree, etc.)
from sklearn.linear_model import LogisticRegression, LinearRegression
from sklearn.tree import DecisionTreeClassifier, DecisionTreeRegressor

# Libraries for ensemble methods and boosting
from sklearn.ensemble import (
    BaggingClassifier,
    RandomForestClassifier,
    AdaBoostClassifier,
    GradientBoostingClassifier,
    StackingClassifier,
)

# Libraries for working with XGBoost
from xgboost import XGBClassifier

# Libraries for oversampling and undersampling (imbalanced data handling)
from imblearn.over_sampling import SMOTE
from imblearn.under_sampling import RandomUnderSampler

# Libraries for scaling and transforming data
from sklearn.preprocessing import LabelEncoder, StandardScaler, MinMaxScaler, OneHotEncoder

# Libraries for missing value imputation
from sklearn.impute import SimpleImputer

# Libraries for creating and managing pipelines
from sklearn.pipeline import Pipeline
from sklearn.compose import ColumnTransformer

# Libraries to help with model tuning and hyperparameter optimization
from sklearn.model_selection import GridSearchCV, RandomizedSearchCV

# Time related functions.
import time

#Imports the tensorflow,keras and layers.
import tensorflow
import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense
from tensorflow.keras.layers import Dense, Input, Dropout,BatchNormalization
from tensorflow.keras import backend

# Google Colab
from google.colab import drive
drive.mount('/content/drive')

Mounted at /content/drive

# Load the Train and Test dataset
train_data = pd.read_csv("/content/drive/MyDrive/Introduction to Neural Networks/ReneWind/Train.csv")
test_data = pd.read_csv("/content/drive/MyDrive/Introduction to Neural Networks/ReneWind/Test.csv")

# Copying Train and Test data to another variable to avoid any changes to original data
df = train_data.copy()
df_test = test_data.copy()

# Display the first few rows of the dataset
print("First few rows of train dataset")
df.head()

First few rows of train dataset

# Display the last few rows of the dataset
print("Last few rows of train dataset")
df.tail()

Last few rows of train dataset

# Display the first few rows of the test dataset
print("First few rows of test dataset")
df_test.head()

First few rows of test dataset

# Display the last few rows of the test dataset
print("Last few rows of test dataset")
df_test.tail()

Last few rows of test dataset

# Display the shape of the train dataset
df.shape
print("There are", df.shape[0], "rows and", df.shape[1], "columns. In the train dataset")

There are 20000 rows and 41 columns. In the train dataset

# Display the data types of the columns in the train dataset
df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 20000 entries, 0 to 19999
Data columns (total 41 columns):
 #   Column  Non-Null Count  Dtype  
---  ------  --------------  -----  
 0   V1      19982 non-null  float64
 1   V2      19982 non-null  float64
 2   V3      20000 non-null  float64
 3   V4      20000 non-null  float64
 4   V5      20000 non-null  float64
 5   V6      20000 non-null  float64
 6   V7      20000 non-null  float64
 7   V8      20000 non-null  float64
 8   V9      20000 non-null  float64
 9   V10     20000 non-null  float64
 10  V11     20000 non-null  float64
 11  V12     20000 non-null  float64
 12  V13     20000 non-null  float64
 13  V14     20000 non-null  float64
 14  V15     20000 non-null  float64
 15  V16     20000 non-null  float64
 16  V17     20000 non-null  float64
 17  V18     20000 non-null  float64
 18  V19     20000 non-null  float64
 19  V20     20000 non-null  float64
 20  V21     20000 non-null  float64
 21  V22     20000 non-null  float64
 22  V23     20000 non-null  float64
 23  V24     20000 non-null  float64
 24  V25     20000 non-null  float64
 25  V26     20000 non-null  float64
 26  V27     20000 non-null  float64
 27  V28     20000 non-null  float64
 28  V29     20000 non-null  float64
 29  V30     20000 non-null  float64
 30  V31     20000 non-null  float64
 31  V32     20000 non-null  float64
 32  V33     20000 non-null  float64
 33  V34     20000 non-null  float64
 34  V35     20000 non-null  float64
 35  V36     20000 non-null  float64
 36  V37     20000 non-null  float64
 37  V38     20000 non-null  float64
 38  V39     20000 non-null  float64
 39  V40     20000 non-null  float64
 40  Target  20000 non-null  int64  
dtypes: float64(40), int64(1)
memory usage: 6.3 MB

# Display the shape of the test dataset
df_test.shape
print("There are", df_test.shape[0], "rows and", df_test.shape[1], "columns. In the test dataset")

There are 5000 rows and 41 columns. In the test dataset

# Display the data types of the columns in the test dataset
df_test.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 5000 entries, 0 to 4999
Data columns (total 41 columns):
 #   Column  Non-Null Count  Dtype  
---  ------  --------------  -----  
 0   V1      4995 non-null   float64
 1   V2      4994 non-null   float64
 2   V3      5000 non-null   float64
 3   V4      5000 non-null   float64
 4   V5      5000 non-null   float64
 5   V6      5000 non-null   float64
 6   V7      5000 non-null   float64
 7   V8      5000 non-null   float64
 8   V9      5000 non-null   float64
 9   V10     5000 non-null   float64
 10  V11     5000 non-null   float64
 11  V12     5000 non-null   float64
 12  V13     5000 non-null   float64
 13  V14     5000 non-null   float64
 14  V15     5000 non-null   float64
 15  V16     5000 non-null   float64
 16  V17     5000 non-null   float64
 17  V18     5000 non-null   float64
 18  V19     5000 non-null   float64
 19  V20     5000 non-null   float64
 20  V21     5000 non-null   float64
 21  V22     5000 non-null   float64
 22  V23     5000 non-null   float64
 23  V24     5000 non-null   float64
 24  V25     5000 non-null   float64
 25  V26     5000 non-null   float64
 26  V27     5000 non-null   float64
 27  V28     5000 non-null   float64
 28  V29     5000 non-null   float64
 29  V30     5000 non-null   float64
 30  V31     5000 non-null   float64
 31  V32     5000 non-null   float64
 32  V33     5000 non-null   float64
 33  V34     5000 non-null   float64
 34  V35     5000 non-null   float64
 35  V36     5000 non-null   float64
 36  V37     5000 non-null   float64
 37  V38     5000 non-null   float64
 38  V39     5000 non-null   float64
 39  V40     5000 non-null   float64
 40  Target  5000 non-null   int64  
dtypes: float64(40), int64(1)
memory usage: 1.6 MB

# Checking missing values across each column of train Data Set
missing_values = df.isnull().sum()
print("The number of missing values on each column of the train data set is:")
missing_values

The number of missing values on each column of the train data set is:

# Checking missing values across each column of test Data Set
missing_values = df_test.isnull().sum()
print("The number of missing values on each column of the test data set is:")
missing_values

The number of missing values on each column of the test data set is:

# Check for complete duplicate records of the train
duplicate_records = df.duplicated().sum()
print("The number of duplicate values on the train data set is:", duplicate_records)

The number of duplicate values on the train data set is: 0

# Check for complete duplicate records of the test
duplicate_records = df_test.duplicated().sum()
print("The number of duplicate values on the test data set is:", duplicate_records)

The number of duplicate values on the test data set is: 0

print("Target Value Counts for train dataset:")
print(df['Target'].value_counts())
print("\nTarget Distribution (%):")
print(df['Target'].value_counts(normalize=True) * 100)

Target Value Counts for train dataset:
Target
0    18890
1     1110
Name: count, dtype: int64

Target Distribution (%):
Target
0   94.4500
1    5.5500
Name: proportion, dtype: float64

print("Target Value Counts for test dataset:")
print(df_test['Target'].value_counts())
print("\nTarget Distribution (%):")
print(df_test['Target'].value_counts(normalize=True) * 100)

Target Value Counts for test dataset:
Target
0    4718
1     282
Name: count, dtype: int64

Target Distribution (%):
Target
0   94.3600
1    5.6400
Name: proportion, dtype: float64

# Display the complete statistical summary for the train dataset
print("Complete Statistical Summary for train dataset")
df.describe(include="all").T

Complete Statistical Summary for train dataset

# Display the complete statistical summary for the test dataset
print("Complete Statistical Summary for test dataset")
df_test.describe(include="all").T

Complete Statistical Summary for test dataset

# Function to plot a boxplot and a histogram along the same scale.
def histogram_boxplot(data, feature, figsize=(12, 7), kde=False, bins=None):
    """
    Boxplot and histogram combined with a title

    data: dataframe
    feature: dataframe column
    figsize: size of figure (default (12,7))
    kde: whether to show the density curve (default False)
    bins: number of bins for histogram (default None)
    """
    f2, (ax_box2, ax_hist2) = plt.subplots(
        nrows=2,  # Number of rows of the subplot grid= 2
        sharex=True,  # x-axis will be shared among all subplots
        gridspec_kw={"height_ratios": (0.25, 0.75)},
        figsize=figsize,
    )  # creating the 2 subplots

    # Add a title to the figure
    f2.suptitle(f"Distribution of {feature}", fontsize=16)

    # Boxplot creation
    sns.boxplot(
        data=data,
        x=feature,
        ax=ax_box2,
        showmeans=True,
        palette="Oranges",
    )

    # Histogram creation
    sns.histplot(
        data=data,
        x=feature,
        kde=kde,
        ax=ax_hist2,
        bins=bins if bins else 50,
        color="orange",
    )

    # Add mean and median lines
    ax_hist2.axvline(data[feature].mean(), color="green", linestyle="--", label="Mean")
    ax_hist2.axvline(
        data[feature].median(), color="black", linestyle="-", label="Median"
    )

    # Add a legend for the mean and median lines
    ax_hist2.legend()

    plt.show()  # Show the plot

# Copying data to another variable to avoid any changes to original data
df_eda = df.copy()

# Selecting all numerical columns ('int64', 'float64')
num_cols = df_eda.select_dtypes(include=["int64", "float64"]).columns

# Iterate through each numerical column and plot the histogram and boxplot combined
for column in num_cols:
    print(f"Distribution of '{column}'")
    print(df_eda[column].describe())
    histogram_boxplot(df_eda, column, bins=50, kde=True)
    print("-" * 100)

Distribution of 'V1'
count   19982.0000
mean       -0.2720
std         3.4416
min       -11.8765
25%        -2.7371
50%        -0.7479
75%         1.8401
max        15.4930
Name: V1, dtype: float64

----------------------------------------------------------------------------------------------------
Distribution of 'V2'
count   19982.0000
mean        0.4404
std         3.1508
min       -12.3200
25%        -1.6407
50%         0.4715
75%         2.5440
max        13.0893
Name: V2, dtype: float64

----------------------------------------------------------------------------------------------------
Distribution of 'V3'
count   20000.0000
mean        2.4847
std         3.3890
min       -10.7081
25%         0.2069
50%         2.2558
75%         4.5662
max        17.0909
Name: V3, dtype: float64

----------------------------------------------------------------------------------------------------
Distribution of 'V4'
count   20000.0000
mean       -0.0832
std         3.4316
min       -15.0821
25%        -2.3477
50%        -0.1352
75%         2.1306
max        13.2364
Name: V4, dtype: float64

----------------------------------------------------------------------------------------------------
Distribution of 'V5'
count   20000.0000
mean       -0.0538
std         2.1048
min        -8.6034
25%        -1.5356
50%        -0.1020
75%         1.3405
max         8.1338
Name: V5, dtype: float64

----------------------------------------------------------------------------------------------------
Distribution of 'V6'
count   20000.0000
mean       -0.9954
std         2.0410
min       -10.2271
25%        -2.3472
50%        -1.0005
75%         0.3803
max         6.9758
Name: V6, dtype: float64

----------------------------------------------------------------------------------------------------
Distribution of 'V7'
count   20000.0000
mean       -0.8793
std         1.7616
min        -7.9497
25%        -2.0309
50%        -0.9172
75%         0.2237
max         8.0061
Name: V7, dtype: float64

----------------------------------------------------------------------------------------------------
Distribution of 'V8'
count   20000.0000
mean       -0.5482
std         3.2958
min       -15.6576
25%        -2.6427
50%        -0.3891
75%         1.7230
max        11.6795
Name: V8, dtype: float64

----------------------------------------------------------------------------------------------------
Distribution of 'V9'
count   20000.0000
mean       -0.0168
std         2.1606
min        -8.5963
25%        -1.4950
50%        -0.0676
75%         1.4092
max         8.1376
Name: V9, dtype: float64

----------------------------------------------------------------------------------------------------
Distribution of 'V10'
count   20000.0000
mean       -0.0130
std         2.1932
min        -9.8540
25%        -1.4112
50%         0.1010
75%         1.4770
max         8.1085
Name: V10, dtype: float64

# Calculate the correlation matrix
corr_matrix = df_eda[num_cols].corr()

# Visual representation of the correlation matrix
plt.figure(figsize=(25, 25))
sns.heatmap(corr_matrix, annot=True, fmt='.2f', cbar=False, cmap='Spectral')
plt.title("Correlation Heatmap")
plt.xlabel("Features")
plt.ylabel("Features")
plt.show()

# Function to create distribution plot
def distribution_plot_wrt_target(data, predictor, target):

    fig, axs = plt.subplots(2, 2, figsize=(12, 10))

    target_uniq = data[target].unique()

    axs[0, 0].set_title(f"Distribution of target for target={str(target_uniq[0])}")
    sns.histplot(
        data=data[data[target] == target_uniq[0]],
        x=predictor,
        kde=True,
        ax=axs[0, 0],
        color="green",
        stat="density",
    )

    axs[0, 1].set_title(f"Distribution of target for target={str(target_uniq[1])}")
    sns.histplot(
        data=data[data[target] == target_uniq[1]],
        x=predictor,
        kde=True,
        ax=axs[0, 1],
        color="blue",
        stat="density",
    )

    axs[1, 0].set_title("Boxplot w.r.t target")
    sns.boxplot(data=data, x=target, y=predictor, ax=axs[1, 0], palette="gist_rainbow")

    axs[1, 1].set_title("Boxplot (without outliers) w.r.t target")
    sns.boxplot(
        data=data,
        x=target,
        y=predictor,
        ax=axs[1, 1],
        showfliers=False,
        palette="gist_rainbow",
    )

    plt.tight_layout()
    plt.show()

# Selecting all numerical columns ('int64', 'float64')
num_cols = df_eda.select_dtypes(include=["int64", "float64"]).columns[:-1]

# Iterate through each numerical column and distribution plot
for column in num_cols:
    distribution_plot_wrt_target(df_eda, column, "Target")
    print("-" * 150)

------------------------------------------------------------------------------------------------------------------------------------------------------

------------------------------------------------------------------------------------------------------------------------------------------------------

------------------------------------------------------------------------------------------------------------------------------------------------------

------------------------------------------------------------------------------------------------------------------------------------------------------

------------------------------------------------------------------------------------------------------------------------------------------------------

------------------------------------------------------------------------------------------------------------------------------------------------------

------------------------------------------------------------------------------------------------------------------------------------------------------

------------------------------------------------------------------------------------------------------------------------------------------------------

------------------------------------------------------------------------------------------------------------------------------------------------------

------------------------------------------------------------------------------------------------------------------------------------------------------

# Copying data to another variable to avoid any changes to original data
df_model = df_eda.copy()

# Outlier detection using boxplot
# Creating a list of numerical columns
numeric_columns = df_model.select_dtypes(include=np.number).columns.tolist()

# Calculate the number of rows needed for the subplots
num_plots = len(num_cols)
num_rows = (num_plots // 3) + (num_plots % 3 > 0)

# Create subplots
plt.figure(figsize=(15, num_rows * 5))
for i, variable in enumerate(num_cols):
    plt.subplot(num_rows, 3, i + 1)
    sns.boxplot(data=df_model, x=variable)
    plt.tight_layout(pad=2)
plt.show()

# Check for outliers in continuous variables
outliers = df_model[num_cols].describe()
print("\nSummary of the numerical features, including outliers:")
display(outliers)

Summary of the numerical features, including outliers:

# Step 1: Define and Prepare the Data
X = df_model.drop("Target", axis=1)
y = df_model["Target"]

display(X.head())
display(y.head())

# Splitting train dataset into training and validation set (75% train, 25% validation), using stratify to maintain target distribution
X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=0.25, random_state=1, stratify=y)

# Checking the number of rows and columns in the X_train data
print("There are", X_train.shape[0], "rows and", X_train.shape[1], "columns in the Training set.", X_train.shape,)

# Checking the number of rows and columns in the X_val data
print("There are", X_val.shape[0], "rows and", X_val.shape[1], "columns in the Test set.", X_val.shape,)

There are 15000 rows and 40 columns in the Training set. (15000, 40)
There are 5000 rows and 40 columns in the Test set. (5000, 40)

print("Percentage of classes in training set:")
print(y_train.value_counts(normalize=True) * 100)

print("")

print("Percentage of classes in validation set:")
print(y_val.value_counts(normalize=True) * 100)

Percentage of classes in training set:
Target
0   94.4533
1    5.5467
Name: proportion, dtype: float64

Percentage of classes in validation set:
Target
0   94.4400
1    5.5600
Name: proportion, dtype: float64

# Dividing test data into X_test and y_test
X_test = df_test.drop(["Target"], axis=1)
y_test = df_test["Target"]

display(X_test.head())
display(y_test.head())

# Checking the number of rows and columns in the X_test data
print("There are", X_test.shape[0], "rows and", X_test.shape[1], "columns in the Test set.", X_test.shape,)

There are 5000 rows and 40 columns in the Test set. (5000, 40)

print("Percentage of classes in test set:")
print(y_test.value_counts(normalize=True) * 100)

Percentage of classes in test set:
Target
0   94.3600
1    5.6400
Name: proportion, dtype: float64

# Creating an instance of the imputer to be used
imputer = SimpleImputer(strategy="median")

# Fit and transform the train data
X_train = pd.DataFrame(imputer.fit_transform(X_train), columns=X_train.columns)

# Transform the validation data without data leakage (using the same imputer trained on X_train)
X_val = pd.DataFrame(imputer.transform(X_val), columns=X_train.columns)

# Transform the test data without data leakage
X_test = pd.DataFrame(imputer.transform(X_test), columns=X_train.columns)

# Checking that no column has missing values in train, validation, or test sets
print("Missing values on the Train dataset")
print(X_train.isna().sum())
print("-" * 50)

print("Missing values on the Validation dataset")
print(X_val.isna().sum())  # Checking missing values in validation set
print("-" * 50)

print("Missing values on the Test dataset")
print(X_test.isna().sum())  # Checking missing values in test set

Missing values on the Train dataset
V1     0
V2     0
V3     0
V4     0
V5     0
V6     0
V7     0
V8     0
V9     0
V10    0
V11    0
V12    0
V13    0
V14    0
V15    0
V16    0
V17    0
V18    0
V19    0
V20    0
V21    0
V22    0
V23    0
V24    0
V25    0
V26    0
V27    0
V28    0
V29    0
V30    0
V31    0
V32    0
V33    0
V34    0
V35    0
V36    0
V37    0
V38    0
V39    0
V40    0
dtype: int64
--------------------------------------------------
Missing values on the Validation dataset
V1     0
V2     0
V3     0
V4     0
V5     0
V6     0
V7     0
V8     0
V9     0
V10    0
V11    0
V12    0
V13    0
V14    0
V15    0
V16    0
V17    0
V18    0
V19    0
V20    0
V21    0
V22    0
V23    0
V24    0
V25    0
V26    0
V27    0
V28    0
V29    0
V30    0
V31    0
V32    0
V33    0
V34    0
V35    0
V36    0
V37    0
V38    0
V39    0
V40    0
dtype: int64
--------------------------------------------------
Missing values on the Test dataset
V1     0
V2     0
V3     0
V4     0
V5     0
V6     0
V7     0
V8     0
V9     0
V10    0
V11    0
V12    0
V13    0
V14    0
V15    0
V16    0
V17    0
V18    0
V19    0
V20    0
V21    0
V22    0
V23    0
V24    0
V25    0
V26    0
V27    0
V28    0
V29    0
V30    0
V31    0
V32    0
V33    0
V34    0
V35    0
V36    0
V37    0
V38    0
V39    0
V40    0
dtype: int64

def plot(history, name):
    """
    Function to plot loss/accuracy

    history: an object which stores the metrics and losses.
    name: can be one of Loss or Accuracy
    """
    fig, ax = plt.subplots() #Creating a subplot with figure and axes.
    plt.plot(history.history[name]) #Plotting the train accuracy or train loss
    plt.plot(history.history['val_'+name]) #Plotting the validation accuracy or validation loss

    plt.title('Model ' + name.capitalize()) #Defining the title of the plot.
    plt.ylabel(name.capitalize()) #Capitalizing the first letter.
    plt.xlabel('Epoch') #Defining the label for the x-axis.
    fig.legend(['Train', 'Validation'], loc="outside right upper") #Defining the legend, loc controls the position of the legend.

# defining a function to compute different metrics to check performance of a classification model built using statsmodels
def model_performance_classification(
    model, predictors, target, threshold=0.5
):
    """
    Function to compute different metrics to check classification model performance

    model: classifier
    predictors: independent variables
    target: dependent variable
    threshold: threshold for classifying the observation as class 1
    """

    # checking which probabilities are greater than threshold
    pred = model.predict(predictors) > threshold
    # pred_temp = model.predict(predictors) > threshold
    # # rounding off the above values to get classes
    # pred = np.round(pred_temp)

    acc = accuracy_score(target, pred)  # to compute Accuracy
    recall = recall_score(target, pred, average='macro')  # to compute Recall
    precision = precision_score(target, pred, average='macro')  # to compute Precision
    f1 = f1_score(target, pred, average='macro')  # to compute F1-score

    # creating a dataframe of metrics
    df_perf = pd.DataFrame(
        {"Accuracy": acc, "Recall": recall, "Precision": precision, "F1 Score": f1,}, index = [0]
    )

    return df_perf

#Defining the columns of the dataframe which are nothing but the hyper parameters and the metrics.
columns = ["# hidden layers","# neurons - hidden layer","activation function - hidden layer","# epochs","batch size","train loss","validation loss","train accuracy","validation accuracy","time (secs)"]

#Creating a pandas dataframe.
results = pd.DataFrame(columns=columns)

# clears the current Keras session, resetting all layers and models previously created, freeing up memory and resources.
tf.keras.backend.clear_session()

#Initializing the neural network
model_0 = Sequential()
model_0.add(Dense(7 ,activation="relu",input_dim=X_train.shape[1]))
model_0.add(Dense(1 ,activation="sigmoid"))

model_0.summary()

Model: "sequential"

┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Layer (type)                    ┃ Output Shape           ┃       Param # ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ dense (Dense)                   │ (None, 7)              │           287 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_1 (Dense)                 │ (None, 1)              │             8 │
└─────────────────────────────────┴────────────────────────┴───────────────┘

 Total params: 295 (1.15 KB)

 Trainable params: 295 (1.15 KB)

 Non-trainable params: 0 (0.00 B)

# defining the batch size and # epochs upfront as we'll be using the same values for all models
epochs = 50
batch_size = 32

optimizer = tf.keras.optimizers.SGD()
model_0.compile(loss='binary_crossentropy', optimizer=optimizer, metrics = ['recall'])

start = time.time()
history = model_0.fit(X_train, y_train, validation_data=(X_val,y_val) , batch_size=batch_size, epochs=epochs)
end=time.time()

Epoch 1/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - loss: 0.3437 - recall: 0.4404 - val_loss: 0.1273 - val_recall: 0.5144
Epoch 2/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.1124 - recall: 0.5566 - val_loss: 0.1125 - val_recall: 0.5612
Epoch 3/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.1012 - recall: 0.6290 - val_loss: 0.1034 - val_recall: 0.5971
Epoch 4/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0915 - recall: 0.6310 - val_loss: 0.0980 - val_recall: 0.6223
Epoch 5/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0889 - recall: 0.6468 - val_loss: 0.0942 - val_recall: 0.6547
Epoch 6/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0812 - recall: 0.6847 - val_loss: 0.0901 - val_recall: 0.6619
Epoch 7/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - loss: 0.0864 - recall: 0.6579 - val_loss: 0.0882 - val_recall: 0.6655
Epoch 8/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - loss: 0.0809 - recall: 0.6764 - val_loss: 0.0853 - val_recall: 0.6799
Epoch 9/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0759 - recall: 0.7105 - val_loss: 0.0834 - val_recall: 0.6978
Epoch 10/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0761 - recall: 0.7368 - val_loss: 0.0802 - val_recall: 0.7122
Epoch 11/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0716 - recall: 0.7146 - val_loss: 0.0784 - val_recall: 0.7194
Epoch 12/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0672 - recall: 0.7672 - val_loss: 0.0754 - val_recall: 0.7662
Epoch 13/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0610 - recall: 0.7735 - val_loss: 0.0732 - val_recall: 0.7806
Epoch 14/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0699 - recall: 0.7594 - val_loss: 0.0719 - val_recall: 0.7842
Epoch 15/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0612 - recall: 0.7752 - val_loss: 0.0701 - val_recall: 0.7842
Epoch 16/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0642 - recall: 0.7599 - val_loss: 0.0693 - val_recall: 0.7986
Epoch 17/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - loss: 0.0623 - recall: 0.7596 - val_loss: 0.0687 - val_recall: 0.8058
Epoch 18/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - loss: 0.0606 - recall: 0.7897 - val_loss: 0.0678 - val_recall: 0.8022
Epoch 19/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - loss: 0.0623 - recall: 0.7575 - val_loss: 0.0674 - val_recall: 0.8058
Epoch 20/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 2ms/step - loss: 0.0640 - recall: 0.7684 - val_loss: 0.0667 - val_recall: 0.8094
Epoch 21/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0515 - recall: 0.8072 - val_loss: 0.0667 - val_recall: 0.8094
Epoch 22/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0586 - recall: 0.7993 - val_loss: 0.0663 - val_recall: 0.8094
Epoch 23/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0627 - recall: 0.7824 - val_loss: 0.0657 - val_recall: 0.8094
Epoch 24/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0471 - recall: 0.8360 - val_loss: 0.0664 - val_recall: 0.8165
Epoch 25/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0568 - recall: 0.8282 - val_loss: 0.0657 - val_recall: 0.8129
Epoch 26/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0510 - recall: 0.8209 - val_loss: 0.0654 - val_recall: 0.8129
Epoch 27/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - loss: 0.0566 - recall: 0.7898 - val_loss: 0.0654 - val_recall: 0.8094
Epoch 28/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - loss: 0.0559 - recall: 0.7941 - val_loss: 0.0649 - val_recall: 0.8058
Epoch 29/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0540 - recall: 0.8113 - val_loss: 0.0646 - val_recall: 0.8129
Epoch 30/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0575 - recall: 0.7959 - val_loss: 0.0653 - val_recall: 0.8165
Epoch 31/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0544 - recall: 0.8150 - val_loss: 0.0648 - val_recall: 0.8129
Epoch 32/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0543 - recall: 0.8200 - val_loss: 0.0648 - val_recall: 0.8129
Epoch 33/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0551 - recall: 0.8018 - val_loss: 0.0646 - val_recall: 0.8094
Epoch 34/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0504 - recall: 0.8140 - val_loss: 0.0647 - val_recall: 0.8201
Epoch 35/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0565 - recall: 0.8173 - val_loss: 0.0647 - val_recall: 0.8094
Epoch 36/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0565 - recall: 0.7958 - val_loss: 0.0647 - val_recall: 0.8165
Epoch 37/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0539 - recall: 0.8036 - val_loss: 0.0649 - val_recall: 0.8165
Epoch 38/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - loss: 0.0550 - recall: 0.8189 - val_loss: 0.0647 - val_recall: 0.8237
Epoch 39/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - loss: 0.0529 - recall: 0.8134 - val_loss: 0.0644 - val_recall: 0.8129
Epoch 40/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0512 - recall: 0.8338 - val_loss: 0.0647 - val_recall: 0.8201
Epoch 41/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0532 - recall: 0.8261 - val_loss: 0.0650 - val_recall: 0.8273
Epoch 42/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0523 - recall: 0.8060 - val_loss: 0.0647 - val_recall: 0.8165
Epoch 43/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0506 - recall: 0.8219 - val_loss: 0.0649 - val_recall: 0.8165
Epoch 44/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0543 - recall: 0.8104 - val_loss: 0.0647 - val_recall: 0.8129
Epoch 45/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0500 - recall: 0.8328 - val_loss: 0.0649 - val_recall: 0.8237
Epoch 46/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0516 - recall: 0.8086 - val_loss: 0.0645 - val_recall: 0.8129
Epoch 47/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0551 - recall: 0.8017 - val_loss: 0.0647 - val_recall: 0.8094
Epoch 48/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - loss: 0.0462 - recall: 0.8208 - val_loss: 0.0641 - val_recall: 0.8165
Epoch 49/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 3s 5ms/step - loss: 0.0550 - recall: 0.8048 - val_loss: 0.0645 - val_recall: 0.8201
Epoch 50/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - loss: 0.0539 - recall: 0.7967 - val_loss: 0.0650 - val_recall: 0.8094

print("Time taken in seconds ",end-start)

Time taken in seconds  69.66447472572327

plot(history, 'loss')

plot(history,'recall')

model_0_train_perf = model_performance_classification(model_0, X_train, y_train)
model_0_train_perf

469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step

model_0_val_perf = model_performance_classification(model_0,X_val,y_val)
model_0_val_perf

157/157 ━━━━━━━━━━━━━━━━━━━━ 0s 1ms/step

y_train_pred_0 = model_0.predict(X_train)
y_val_pred_0 = model_0.predict(X_val)

469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 1ms/step
157/157 ━━━━━━━━━━━━━━━━━━━━ 0s 1ms/step

print("Classification Report - Train data Model_0",end="\n\n")
cr_train_model_0 = classification_report(y_train,y_train_pred_0>0.5)
print(cr_train_model_0)

Classification Report - Train data Model_0

              precision    recall  f1-score   support

           0       0.99      1.00      0.99     14168
           1       0.96      0.81      0.88       832

    accuracy                           0.99     15000
   macro avg       0.97      0.90      0.94     15000
weighted avg       0.99      0.99      0.99     15000

print("Classification Report - Validation data Model_0",end="\n\n")
cr_val_model_0 = classification_report(y_val,y_val_pred_0>0.5)
print(cr_val_model_0)

Classification Report - Validation data Model_0

              precision    recall  f1-score   support

           0       0.99      1.00      0.99      4722
           1       0.97      0.81      0.88       278

    accuracy                           0.99      5000
   macro avg       0.98      0.90      0.94      5000
weighted avg       0.99      0.99      0.99      5000

# clears the current Keras session, resetting all layers and models previously created, freeing up memory and resources.
tf.keras.backend.clear_session()

#Initializing the neural network
model_1 = Sequential()
model_1.add(Dense(128 ,activation="relu",input_dim=X_train.shape[1]))
model_1.add(Dense(7 ,activation="relu"))
model_1.add(Dense(1 ,activation="sigmoid"))

model_1.summary()

Model: "sequential"

┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Layer (type)                    ┃ Output Shape           ┃       Param # ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ dense (Dense)                   │ (None, 128)            │         5,248 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_1 (Dense)                 │ (None, 7)              │           903 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_2 (Dense)                 │ (None, 1)              │             8 │
└─────────────────────────────────┴────────────────────────┴───────────────┘

 Total params: 6,159 (24.06 KB)

 Trainable params: 6,159 (24.06 KB)

 Non-trainable params: 0 (0.00 B)

# Setting training configuration
epochs = 50
batch_size = 64

optimizer = tf.keras.optimizers.SGD()
model_1.compile(loss='binary_crossentropy', optimizer=optimizer, metrics = ['recall'])

start = time.time()
history = model_1.fit(X_train, y_train, validation_data=(X_val,y_val) , batch_size=batch_size, epochs=epochs)
end=time.time()

Epoch 1/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 3s 8ms/step - loss: 0.2427 - recall: 0.2474 - val_loss: 0.1124 - val_recall: 0.5180
Epoch 2/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 2s 8ms/step - loss: 0.0996 - recall: 0.5248 - val_loss: 0.0945 - val_recall: 0.6511
Epoch 3/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 3s 9ms/step - loss: 0.0948 - recall: 0.6314 - val_loss: 0.0862 - val_recall: 0.6799
Epoch 4/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 2s 6ms/step - loss: 0.0736 - recall: 0.6957 - val_loss: 0.0805 - val_recall: 0.7374
Epoch 5/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 5ms/step - loss: 0.0712 - recall: 0.7633 - val_loss: 0.0760 - val_recall: 0.7518
Epoch 6/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - loss: 0.0667 - recall: 0.7511 - val_loss: 0.0727 - val_recall: 0.7842
Epoch 7/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0639 - recall: 0.7863 - val_loss: 0.0704 - val_recall: 0.7806
Epoch 8/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0635 - recall: 0.7907 - val_loss: 0.0682 - val_recall: 0.7842
Epoch 9/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0682 - recall: 0.7926 - val_loss: 0.0670 - val_recall: 0.7914
Epoch 10/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0568 - recall: 0.8021 - val_loss: 0.0662 - val_recall: 0.7950
Epoch 11/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0538 - recall: 0.8357 - val_loss: 0.0649 - val_recall: 0.8022
Epoch 12/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0573 - recall: 0.8112 - val_loss: 0.0643 - val_recall: 0.8058
Epoch 13/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0500 - recall: 0.8531 - val_loss: 0.0630 - val_recall: 0.8094
Epoch 14/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0528 - recall: 0.8359 - val_loss: 0.0624 - val_recall: 0.8165
Epoch 15/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0515 - recall: 0.8261 - val_loss: 0.0622 - val_recall: 0.8058
Epoch 16/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0500 - recall: 0.8525 - val_loss: 0.0613 - val_recall: 0.8309
Epoch 17/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0465 - recall: 0.8574 - val_loss: 0.0609 - val_recall: 0.8273
Epoch 18/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0526 - recall: 0.8452 - val_loss: 0.0603 - val_recall: 0.8237
Epoch 19/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0503 - recall: 0.8617 - val_loss: 0.0601 - val_recall: 0.8309
Epoch 20/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 4ms/step - loss: 0.0462 - recall: 0.8583 - val_loss: 0.0605 - val_recall: 0.8201
Epoch 21/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 4ms/step - loss: 0.0431 - recall: 0.8592 - val_loss: 0.0597 - val_recall: 0.8345
Epoch 22/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 4ms/step - loss: 0.0470 - recall: 0.8612 - val_loss: 0.0591 - val_recall: 0.8345
Epoch 23/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 4ms/step - loss: 0.0504 - recall: 0.8350 - val_loss: 0.0594 - val_recall: 0.8345
Epoch 24/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0470 - recall: 0.8639 - val_loss: 0.0591 - val_recall: 0.8345
Epoch 25/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0426 - recall: 0.8626 - val_loss: 0.0588 - val_recall: 0.8417
Epoch 26/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0421 - recall: 0.8751 - val_loss: 0.0587 - val_recall: 0.8345
Epoch 27/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0432 - recall: 0.8637 - val_loss: 0.0584 - val_recall: 0.8417
Epoch 28/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - loss: 0.0419 - recall: 0.8617 - val_loss: 0.0582 - val_recall: 0.8453
Epoch 29/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0393 - recall: 0.8930 - val_loss: 0.0580 - val_recall: 0.8381
Epoch 30/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0448 - recall: 0.8600 - val_loss: 0.0585 - val_recall: 0.8525
Epoch 31/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0363 - recall: 0.8851 - val_loss: 0.0580 - val_recall: 0.8381
Epoch 32/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0426 - recall: 0.8602 - val_loss: 0.0580 - val_recall: 0.8453
Epoch 33/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0347 - recall: 0.9023 - val_loss: 0.0578 - val_recall: 0.8489
Epoch 34/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0352 - recall: 0.8885 - val_loss: 0.0578 - val_recall: 0.8525
Epoch 35/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0429 - recall: 0.8643 - val_loss: 0.0575 - val_recall: 0.8453
Epoch 36/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0403 - recall: 0.8762 - val_loss: 0.0572 - val_recall: 0.8453
Epoch 37/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0407 - recall: 0.8720 - val_loss: 0.0569 - val_recall: 0.8417
Epoch 38/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0442 - recall: 0.8700 - val_loss: 0.0570 - val_recall: 0.8453
Epoch 39/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 4ms/step - loss: 0.0414 - recall: 0.8747 - val_loss: 0.0568 - val_recall: 0.8453
Epoch 40/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 5ms/step - loss: 0.0378 - recall: 0.8911 - val_loss: 0.0569 - val_recall: 0.8453
Epoch 41/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0401 - recall: 0.8792 - val_loss: 0.0567 - val_recall: 0.8489
Epoch 42/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0444 - recall: 0.8830 - val_loss: 0.0571 - val_recall: 0.8489
Epoch 43/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0442 - recall: 0.8762 - val_loss: 0.0571 - val_recall: 0.8417
Epoch 44/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0340 - recall: 0.8856 - val_loss: 0.0564 - val_recall: 0.8417
Epoch 45/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0434 - recall: 0.8662 - val_loss: 0.0568 - val_recall: 0.8417
Epoch 46/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0407 - recall: 0.8773 - val_loss: 0.0569 - val_recall: 0.8525
Epoch 47/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0380 - recall: 0.8861 - val_loss: 0.0564 - val_recall: 0.8561
Epoch 48/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0414 - recall: 0.8785 - val_loss: 0.0566 - val_recall: 0.8489
Epoch 49/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0381 - recall: 0.8877 - val_loss: 0.0565 - val_recall: 0.8453
Epoch 50/50
235/235 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - loss: 0.0340 - recall: 0.9034 - val_loss: 0.0562 - val_recall: 0.8561

print("Time taken in seconds ",end-start)

Time taken in seconds  47.98444724082947

plot(history,'loss')

plot(history,'recall')

model_1_train_perf = model_performance_classification(model_1,X_train,y_train)
model_1_train_perf

469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step

model_1_val_perf = model_performance_classification(model_1,X_val,y_val)
model_1_val_perf

157/157 ━━━━━━━━━━━━━━━━━━━━ 0s 1ms/step

y_train_pred_1 = model_1.predict(X_train)
y_val_pred_1 = model_1.predict(X_val)

469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 1ms/step
157/157 ━━━━━━━━━━━━━━━━━━━━ 0s 1ms/step

print("Classification Report - Train data Model_1", end="\n\n")
cr_train_model_1 = classification_report(y_train,y_train_pred_1 > 0.5)
print(cr_train_model_1)

Classification Report - Train data Model_1

              precision    recall  f1-score   support

           0       0.99      1.00      1.00     14168
           1       0.99      0.89      0.93       832

    accuracy                           0.99     15000
   macro avg       0.99      0.94      0.97     15000
weighted avg       0.99      0.99      0.99     15000

print("Classification Report - Validation data Model_1", end="\n\n")
cr_val_model_1 = classification_report(y_val,y_val_pred_1 > 0.5)
print(cr_val_model_1)

Classification Report - Validation data Model_1

              precision    recall  f1-score   support

           0       0.99      1.00      1.00      4722
           1       0.97      0.86      0.91       278

    accuracy                           0.99      5000
   macro avg       0.98      0.93      0.95      5000
weighted avg       0.99      0.99      0.99      5000

# clears the current Keras session, resetting all layers and models previously created, freeing up memory and resources.
tf.keras.backend.clear_session()

#Initializing the neural network
from tensorflow.keras.layers import Dropout
model_2 = Sequential()
model_2.add(Dense(16 ,activation="relu",input_dim=X_train.shape[1]))
model_2.add(Dropout(0.5))
model_2.add(Dense(128,activation = "relu"))
model_2.add(Dense(64,activation = "tanh"))
model_2.add(Dense(1,activation="sigmoid"))

model_2.summary()

Model: "sequential"

┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Layer (type)                    ┃ Output Shape           ┃       Param # ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ dense (Dense)                   │ (None, 16)             │           656 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dropout (Dropout)               │ (None, 16)             │             0 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_1 (Dense)                 │ (None, 128)            │         2,176 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_2 (Dense)                 │ (None, 64)             │         8,256 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_3 (Dense)                 │ (None, 1)              │            65 │
└─────────────────────────────────┴────────────────────────┴───────────────┘

 Total params: 11,153 (43.57 KB)

 Trainable params: 11,153 (43.57 KB)

 Non-trainable params: 0 (0.00 B)

# Setting training configuration
epochs = 50
batch_size = 32

optimizer = tf.keras.optimizers.SGD()
model_2.compile(loss='binary_crossentropy', optimizer='sgd', metrics = ['Recall'])

start = time.time()
history = model_2.fit(X_train, y_train, validation_data=(X_val,y_val) , batch_size=batch_size, epochs=epochs)
end=time.time()

Epoch 1/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - Recall: 3.2082e-04 - loss: 0.2584 - val_Recall: 0.0000e+00 - val_loss: 0.1596
Epoch 2/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.0257 - loss: 0.1812 - val_Recall: 0.1727 - val_loss: 0.1353
Epoch 3/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.1632 - loss: 0.1687 - val_Recall: 0.3561 - val_loss: 0.1155
Epoch 4/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.2928 - loss: 0.1354 - val_Recall: 0.4604 - val_loss: 0.1029
Epoch 5/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.3660 - loss: 0.1342 - val_Recall: 0.5072 - val_loss: 0.0966
Epoch 6/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 3s 4ms/step - Recall: 0.4254 - loss: 0.1180 - val_Recall: 0.5899 - val_loss: 0.0902
Epoch 7/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.4646 - loss: 0.1133 - val_Recall: 0.6223 - val_loss: 0.0866
Epoch 8/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.4403 - loss: 0.1104 - val_Recall: 0.6583 - val_loss: 0.0825
Epoch 9/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.4833 - loss: 0.1086 - val_Recall: 0.6691 - val_loss: 0.0805
Epoch 10/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.4992 - loss: 0.1093 - val_Recall: 0.6583 - val_loss: 0.0806
Epoch 11/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.4821 - loss: 0.1060 - val_Recall: 0.7014 - val_loss: 0.0769
Epoch 12/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.4950 - loss: 0.1051 - val_Recall: 0.7050 - val_loss: 0.0750
Epoch 13/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5098 - loss: 0.1088 - val_Recall: 0.7050 - val_loss: 0.0746
Epoch 14/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - Recall: 0.5302 - loss: 0.0963 - val_Recall: 0.7158 - val_loss: 0.0737
Epoch 15/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 3s 4ms/step - Recall: 0.5234 - loss: 0.0990 - val_Recall: 0.7302 - val_loss: 0.0719
Epoch 16/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5206 - loss: 0.0945 - val_Recall: 0.7122 - val_loss: 0.0720
Epoch 17/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5727 - loss: 0.0919 - val_Recall: 0.7266 - val_loss: 0.0718
Epoch 18/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5719 - loss: 0.0908 - val_Recall: 0.7194 - val_loss: 0.0715
Epoch 19/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5776 - loss: 0.0905 - val_Recall: 0.7482 - val_loss: 0.0709
Epoch 20/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5488 - loss: 0.1010 - val_Recall: 0.7374 - val_loss: 0.0702
Epoch 21/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5808 - loss: 0.0867 - val_Recall: 0.7626 - val_loss: 0.0693
Epoch 22/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 3s 3ms/step - Recall: 0.5932 - loss: 0.0847 - val_Recall: 0.7626 - val_loss: 0.0688
Epoch 23/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.5821 - loss: 0.0898 - val_Recall: 0.7662 - val_loss: 0.0689
Epoch 24/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - Recall: 0.5918 - loss: 0.0817 - val_Recall: 0.7446 - val_loss: 0.0685
Epoch 25/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5867 - loss: 0.0829 - val_Recall: 0.7662 - val_loss: 0.0678
Epoch 26/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5990 - loss: 0.0844 - val_Recall: 0.7626 - val_loss: 0.0671
Epoch 27/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5748 - loss: 0.0874 - val_Recall: 0.7806 - val_loss: 0.0676
Epoch 28/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5874 - loss: 0.0889 - val_Recall: 0.7662 - val_loss: 0.0669
Epoch 29/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5962 - loss: 0.0863 - val_Recall: 0.7554 - val_loss: 0.0670
Epoch 30/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5924 - loss: 0.0827 - val_Recall: 0.7806 - val_loss: 0.0662
Epoch 31/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5676 - loss: 0.0828 - val_Recall: 0.7842 - val_loss: 0.0656
Epoch 32/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.5906 - loss: 0.0842 - val_Recall: 0.7770 - val_loss: 0.0659
Epoch 33/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.5788 - loss: 0.0865 - val_Recall: 0.7626 - val_loss: 0.0663
Epoch 34/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5862 - loss: 0.0792 - val_Recall: 0.7734 - val_loss: 0.0653
Epoch 35/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5554 - loss: 0.0878 - val_Recall: 0.7662 - val_loss: 0.0658
Epoch 36/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.6388 - loss: 0.0845 - val_Recall: 0.7698 - val_loss: 0.0660
Epoch 37/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5914 - loss: 0.0791 - val_Recall: 0.7698 - val_loss: 0.0657
Epoch 38/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.6096 - loss: 0.0777 - val_Recall: 0.7698 - val_loss: 0.0658
Epoch 39/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.6274 - loss: 0.0776 - val_Recall: 0.7590 - val_loss: 0.0647
Epoch 40/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.6008 - loss: 0.0745 - val_Recall: 0.7662 - val_loss: 0.0650
Epoch 41/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.5825 - loss: 0.0813 - val_Recall: 0.7662 - val_loss: 0.0656
Epoch 42/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.5774 - loss: 0.0827 - val_Recall: 0.7734 - val_loss: 0.0655
Epoch 43/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.6380 - loss: 0.0774 - val_Recall: 0.7698 - val_loss: 0.0644
Epoch 44/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.6045 - loss: 0.0733 - val_Recall: 0.7770 - val_loss: 0.0649
Epoch 45/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5778 - loss: 0.0804 - val_Recall: 0.7770 - val_loss: 0.0647
Epoch 46/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5978 - loss: 0.0795 - val_Recall: 0.7878 - val_loss: 0.0650
Epoch 47/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 3s 3ms/step - Recall: 0.5832 - loss: 0.0784 - val_Recall: 0.7662 - val_loss: 0.0648
Epoch 48/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.5862 - loss: 0.0848 - val_Recall: 0.7734 - val_loss: 0.0652
Epoch 49/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.6300 - loss: 0.0722 - val_Recall: 0.7626 - val_loss: 0.0657
Epoch 50/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 5ms/step - Recall: 0.6016 - loss: 0.0741 - val_Recall: 0.7734 - val_loss: 0.0646

print("Time taken in seconds ",end-start)

Time taken in seconds  77.34527611732483

plot(history,'loss')

plot(history,'Recall')

model_2_train_perf = model_performance_classification(model_2,X_train,y_train)
model_2_train_perf

469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 1ms/step

model_2_val_perf = model_performance_classification(model_2,X_val,y_val)
model_2_val_perf

157/157 ━━━━━━━━━━━━━━━━━━━━ 0s 1ms/step

y_train_pred_2 = model_2.predict(X_train)
y_val_pred_2 = model_2.predict(X_val)

469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 1ms/step
157/157 ━━━━━━━━━━━━━━━━━━━━ 0s 2ms/step

print("Classification Report - Train data Model_2", end="\n\n")
cr_train_model_2 = classification_report(y_train,y_train_pred_2 > 0.5)
print(cr_train_model_2)

Classification Report - Train data Model_2

              precision    recall  f1-score   support

           0       0.99      1.00      0.99     14168
           1       0.98      0.74      0.85       832

    accuracy                           0.98     15000
   macro avg       0.98      0.87      0.92     15000
weighted avg       0.98      0.98      0.98     15000

print("Classification Report - Validation data Model_2", end="\n\n")
cr_val_model_2 = classification_report(y_val , y_val_pred_2 > 0.5)
print(cr_val_model_2)

Classification Report - Validation data Model_2

              precision    recall  f1-score   support

           0       0.99      1.00      0.99      4722
           1       0.96      0.77      0.86       278

    accuracy                           0.99      5000
   macro avg       0.97      0.89      0.92      5000
weighted avg       0.99      0.99      0.98      5000

# Calculate class weights for imbalanced dataset
cw = (y_train.shape[0]) / np.bincount(y_train.astype(int)) # Convert y_train to integers

# Create a dictionary mapping class indices to their respective class weights
cw_dict = {}
for i in range(cw.shape[0]):
    cw_dict[i] = cw[i]

cw_dict

{0: np.float64(1.0587238848108413), 1: np.float64(18.028846153846153)}

# clears the current Keras session, resetting all layers and models previously created, freeing up memory and resources.
tf.keras.backend.clear_session()

model_3 = Sequential()
model_3.add(Dense(16 ,activation="relu",input_dim=X_train.shape[1]))
model_3.add(Dropout(0.5))
model_3.add(Dense(128 ,activation="relu"))
model_3.add(Dense(64 , activation = "tanh"))
model_3.add(Dense(1,activation="sigmoid"))

model_3.summary()

Model: "sequential"

┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Layer (type)                    ┃ Output Shape           ┃       Param # ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ dense (Dense)                   │ (None, 16)             │           656 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dropout (Dropout)               │ (None, 16)             │             0 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_1 (Dense)                 │ (None, 128)            │         2,176 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_2 (Dense)                 │ (None, 64)             │         8,256 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_3 (Dense)                 │ (None, 1)              │            65 │
└─────────────────────────────────┴────────────────────────┴───────────────┘

 Total params: 11,153 (43.57 KB)

 Trainable params: 11,153 (43.57 KB)

 Non-trainable params: 0 (0.00 B)

# Set training parameters
epochs = 50
batch_size = 32

optimizer = tf.keras.optimizers.SGD()
model_3.compile(loss='binary_crossentropy', optimizer="sgd", metrics = ['Recall'])

start = time.time()
history = model_3.fit(X_train, y_train, validation_data=(X_val,y_val) , batch_size=batch_size, epochs=epochs,class_weight=cw_dict)
end=time.time()

Epoch 1/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - Recall: 0.7750 - loss: 1.0652 - val_Recall: 0.8777 - val_loss: 0.3157
Epoch 2/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8256 - loss: 0.6728 - val_Recall: 0.8777 - val_loss: 0.3066
Epoch 3/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8126 - loss: 0.6605 - val_Recall: 0.8741 - val_loss: 0.2608
Epoch 4/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8397 - loss: 0.5814 - val_Recall: 0.8813 - val_loss: 0.2573
Epoch 5/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - Recall: 0.8319 - loss: 0.6040 - val_Recall: 0.8705 - val_loss: 0.2341
Epoch 6/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.8391 - loss: 0.5825 - val_Recall: 0.8885 - val_loss: 0.2366
Epoch 7/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.8072 - loss: 0.6055 - val_Recall: 0.8705 - val_loss: 0.1801
Epoch 8/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8215 - loss: 0.5811 - val_Recall: 0.8777 - val_loss: 0.1746
Epoch 9/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8398 - loss: 0.5684 - val_Recall: 0.8777 - val_loss: 0.1732
Epoch 10/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8333 - loss: 0.5679 - val_Recall: 0.8741 - val_loss: 0.1741
Epoch 11/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8297 - loss: 0.5199 - val_Recall: 0.8741 - val_loss: 0.1799
Epoch 12/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8540 - loss: 0.5300 - val_Recall: 0.8849 - val_loss: 0.1711
Epoch 13/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8383 - loss: 0.5681 - val_Recall: 0.8849 - val_loss: 0.1754
Epoch 14/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8438 - loss: 0.5169 - val_Recall: 0.8741 - val_loss: 0.1854
Epoch 15/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - Recall: 0.8497 - loss: 0.5254 - val_Recall: 0.8741 - val_loss: 0.1753
Epoch 16/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 3s 4ms/step - Recall: 0.8641 - loss: 0.4634 - val_Recall: 0.8777 - val_loss: 0.1553
Epoch 17/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8658 - loss: 0.4888 - val_Recall: 0.8993 - val_loss: 0.3066
Epoch 18/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8470 - loss: 0.5317 - val_Recall: 0.8921 - val_loss: 0.1890
Epoch 19/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8440 - loss: 0.5213 - val_Recall: 0.8669 - val_loss: 0.1490
Epoch 20/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8440 - loss: 0.5255 - val_Recall: 0.8849 - val_loss: 0.1928
Epoch 21/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8506 - loss: 0.4903 - val_Recall: 0.8777 - val_loss: 0.1655
Epoch 22/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8578 - loss: 0.5063 - val_Recall: 0.8849 - val_loss: 0.2571
Epoch 23/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8602 - loss: 0.5010 - val_Recall: 0.8813 - val_loss: 0.2446
Epoch 24/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - Recall: 0.8589 - loss: 0.4591 - val_Recall: 0.8849 - val_loss: 0.2090
Epoch 25/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.8615 - loss: 0.5053 - val_Recall: 0.8849 - val_loss: 0.1685
Epoch 26/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.8540 - loss: 0.5017 - val_Recall: 0.8777 - val_loss: 0.1414
Epoch 27/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8535 - loss: 0.4807 - val_Recall: 0.8849 - val_loss: 0.1703
Epoch 28/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8641 - loss: 0.5025 - val_Recall: 0.8741 - val_loss: 0.1209
Epoch 29/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8463 - loss: 0.5142 - val_Recall: 0.8777 - val_loss: 0.1673
Epoch 30/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8592 - loss: 0.4733 - val_Recall: 0.8777 - val_loss: 0.1452
Epoch 31/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8556 - loss: 0.4779 - val_Recall: 0.8777 - val_loss: 0.1503
Epoch 32/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8537 - loss: 0.5168 - val_Recall: 0.8813 - val_loss: 0.1601
Epoch 33/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8728 - loss: 0.4638 - val_Recall: 0.8849 - val_loss: 0.1424
Epoch 34/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.8206 - loss: 0.5561 - val_Recall: 0.8813 - val_loss: 0.1926
Epoch 35/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.8683 - loss: 0.4765 - val_Recall: 0.8849 - val_loss: 0.1870
Epoch 36/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - Recall: 0.8431 - loss: 0.4952 - val_Recall: 0.8777 - val_loss: 0.2281
Epoch 37/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8574 - loss: 0.5077 - val_Recall: 0.8777 - val_loss: 0.1259
Epoch 38/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8522 - loss: 0.4944 - val_Recall: 0.8849 - val_loss: 0.2105
Epoch 39/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 3s 3ms/step - Recall: 0.8529 - loss: 0.5204 - val_Recall: 0.8813 - val_loss: 0.1649
Epoch 40/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8453 - loss: 0.5037 - val_Recall: 0.8813 - val_loss: 0.2207
Epoch 41/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8374 - loss: 0.5355 - val_Recall: 0.8813 - val_loss: 0.1753
Epoch 42/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - Recall: 0.8607 - loss: 0.5025 - val_Recall: 0.8777 - val_loss: 0.1538
Epoch 43/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 3s 4ms/step - Recall: 0.8477 - loss: 0.4973 - val_Recall: 0.8813 - val_loss: 0.1853
Epoch 44/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8722 - loss: 0.4854 - val_Recall: 0.8813 - val_loss: 0.1614
Epoch 45/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8627 - loss: 0.4681 - val_Recall: 0.8741 - val_loss: 0.1542
Epoch 46/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8591 - loss: 0.4948 - val_Recall: 0.8705 - val_loss: 0.1154
Epoch 47/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8149 - loss: 0.5637 - val_Recall: 0.8777 - val_loss: 0.1791
Epoch 48/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8592 - loss: 0.4865 - val_Recall: 0.8813 - val_loss: 0.1979
Epoch 49/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8458 - loss: 0.4863 - val_Recall: 0.8741 - val_loss: 0.1725
Epoch 50/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8394 - loss: 0.4926 - val_Recall: 0.8813 - val_loss: 0.1627

print("Time taken in seconds ",end-start)

Time taken in seconds  74.89451193809509

plot(history,'loss')

plot(history,'Recall')

model_3_train_perf = model_performance_classification(model_3,X_train,y_train)
model_3_train_perf

469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step

model_3_val_perf = model_performance_classification(model_3,X_val,y_val)
model_3_val_perf

157/157 ━━━━━━━━━━━━━━━━━━━━ 0s 1ms/step

y_train_pred_3 = model_3.predict(X_train)
y_val_pred_3 = model_3.predict(X_val)

469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 1ms/step
157/157 ━━━━━━━━━━━━━━━━━━━━ 0s 1ms/step

print("Classification Report - Train data Model_3", end="\n\n")
cr_train_model_3 = classification_report(y_train,y_train_pred_3 > 0.5)
print(cr_train_model_3)

Classification Report - Train data Model_3

              precision    recall  f1-score   support

           0       0.99      0.97      0.98     14168
           1       0.66      0.91      0.76       832

    accuracy                           0.97     15000
   macro avg       0.83      0.94      0.87     15000
weighted avg       0.98      0.97      0.97     15000

print("Classification Report - Validation data Model_3", end="\n\n")
cr_val_model_3 = classification_report(y_val,y_val_pred_3 > 0.5)
print(cr_val_model_3)

Classification Report - Validation data Model_3

              precision    recall  f1-score   support

           0       0.99      0.97      0.98      4722
           1       0.63      0.88      0.73       278

    accuracy                           0.96      5000
   macro avg       0.81      0.93      0.86      5000
weighted avg       0.97      0.96      0.97      5000

# clears the current Keras session, resetting all layers and models previously created, freeing up memory and resources.
tf.keras.backend.clear_session()

#Initializing the neural network
model_4 = Sequential()
model_4.add(Dense(128 ,activation="relu",input_dim=X_train.shape[1]))
model_4.add(Dense(64 ,activation="tanh"))
model_4.add(Dense(32 ,activation="tanh"))
model_4.add(Dense(1 ,activation="sigmoid"))

model_4.summary()

Model: "sequential"

┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Layer (type)                    ┃ Output Shape           ┃       Param # ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ dense (Dense)                   │ (None, 128)            │         5,248 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_1 (Dense)                 │ (None, 64)             │         8,256 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_2 (Dense)                 │ (None, 32)             │         2,080 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_3 (Dense)                 │ (None, 1)              │            33 │
└─────────────────────────────────┴────────────────────────┴───────────────┘

 Total params: 15,617 (61.00 KB)

 Trainable params: 15,617 (61.00 KB)

 Non-trainable params: 0 (0.00 B)

# Set training parameters
epochs = 50
batch_size = 32

optimizer = tf.keras.optimizers.Adam()
model_4.compile(loss='binary_crossentropy', optimizer=optimizer, metrics = ['Recall'])

start = time.time()
history = model_4.fit(X_train, y_train, validation_data=(X_val,y_val) , batch_size=batch_size, epochs=epochs)
end=time.time()

Epoch 1/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 4s 4ms/step - Recall: 0.5219 - loss: 0.1082 - val_Recall: 0.8453 - val_loss: 0.0509
Epoch 2/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 5ms/step - Recall: 0.8677 - loss: 0.0419 - val_Recall: 0.8453 - val_loss: 0.0535
Epoch 3/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.8747 - loss: 0.0398 - val_Recall: 0.8453 - val_loss: 0.0468
Epoch 4/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8873 - loss: 0.0362 - val_Recall: 0.8489 - val_loss: 0.0485
Epoch 5/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8888 - loss: 0.0351 - val_Recall: 0.8777 - val_loss: 0.0470
Epoch 6/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9055 - loss: 0.0311 - val_Recall: 0.8597 - val_loss: 0.0478
Epoch 7/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9048 - loss: 0.0293 - val_Recall: 0.8741 - val_loss: 0.0446
Epoch 8/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8997 - loss: 0.0286 - val_Recall: 0.8885 - val_loss: 0.0463
Epoch 9/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9142 - loss: 0.0291 - val_Recall: 0.8885 - val_loss: 0.0461
Epoch 10/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.9018 - loss: 0.0305 - val_Recall: 0.8741 - val_loss: 0.0472
Epoch 11/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 5ms/step - Recall: 0.9026 - loss: 0.0310 - val_Recall: 0.8561 - val_loss: 0.0466
Epoch 12/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - Recall: 0.9150 - loss: 0.0258 - val_Recall: 0.8777 - val_loss: 0.0456
Epoch 13/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9191 - loss: 0.0259 - val_Recall: 0.8741 - val_loss: 0.0477
Epoch 14/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9186 - loss: 0.0238 - val_Recall: 0.8777 - val_loss: 0.0484
Epoch 15/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9074 - loss: 0.0264 - val_Recall: 0.8849 - val_loss: 0.0467
Epoch 16/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9271 - loss: 0.0211 - val_Recall: 0.8813 - val_loss: 0.0501
Epoch 17/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9152 - loss: 0.0252 - val_Recall: 0.8741 - val_loss: 0.0470
Epoch 18/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9158 - loss: 0.0252 - val_Recall: 0.8561 - val_loss: 0.0506
Epoch 19/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 5ms/step - Recall: 0.9084 - loss: 0.0250 - val_Recall: 0.8705 - val_loss: 0.0502
Epoch 20/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 5ms/step - Recall: 0.9136 - loss: 0.0254 - val_Recall: 0.8885 - val_loss: 0.0519
Epoch 21/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9263 - loss: 0.0208 - val_Recall: 0.8813 - val_loss: 0.0512
Epoch 22/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9192 - loss: 0.0235 - val_Recall: 0.8705 - val_loss: 0.0527
Epoch 23/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9101 - loss: 0.0232 - val_Recall: 0.8633 - val_loss: 0.0533
Epoch 24/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9128 - loss: 0.0212 - val_Recall: 0.8885 - val_loss: 0.0569
Epoch 25/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9396 - loss: 0.0189 - val_Recall: 0.8813 - val_loss: 0.0527
Epoch 26/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9222 - loss: 0.0187 - val_Recall: 0.8813 - val_loss: 0.0487
Epoch 27/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9357 - loss: 0.0164 - val_Recall: 0.8741 - val_loss: 0.0539
Epoch 28/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.9306 - loss: 0.0177 - val_Recall: 0.8741 - val_loss: 0.0523
Epoch 29/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 5ms/step - Recall: 0.9142 - loss: 0.0191 - val_Recall: 0.8885 - val_loss: 0.0537
Epoch 30/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - Recall: 0.9155 - loss: 0.0180 - val_Recall: 0.8669 - val_loss: 0.0597
Epoch 31/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9204 - loss: 0.0200 - val_Recall: 0.8849 - val_loss: 0.0548
Epoch 32/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9225 - loss: 0.0182 - val_Recall: 0.8777 - val_loss: 0.0566
Epoch 33/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9530 - loss: 0.0121 - val_Recall: 0.8633 - val_loss: 0.0599
Epoch 34/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9186 - loss: 0.0196 - val_Recall: 0.8813 - val_loss: 0.0586
Epoch 35/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9414 - loss: 0.0148 - val_Recall: 0.8741 - val_loss: 0.0569
Epoch 36/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9377 - loss: 0.0144 - val_Recall: 0.8705 - val_loss: 0.0567
Epoch 37/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 5ms/step - Recall: 0.9379 - loss: 0.0145 - val_Recall: 0.8633 - val_loss: 0.0573
Epoch 38/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.9370 - loss: 0.0142 - val_Recall: 0.8597 - val_loss: 0.0614
Epoch 39/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9233 - loss: 0.0164 - val_Recall: 0.8777 - val_loss: 0.0570
Epoch 40/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9242 - loss: 0.0198 - val_Recall: 0.8813 - val_loss: 0.0569
Epoch 41/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9474 - loss: 0.0132 - val_Recall: 0.8777 - val_loss: 0.0563
Epoch 42/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9446 - loss: 0.0124 - val_Recall: 0.8741 - val_loss: 0.0588
Epoch 43/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9406 - loss: 0.0144 - val_Recall: 0.8777 - val_loss: 0.0633
Epoch 44/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9501 - loss: 0.0127 - val_Recall: 0.8633 - val_loss: 0.0593
Epoch 45/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.9414 - loss: 0.0104 - val_Recall: 0.8561 - val_loss: 0.0625
Epoch 46/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 5ms/step - Recall: 0.9426 - loss: 0.0117 - val_Recall: 0.8777 - val_loss: 0.0633
Epoch 47/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.9471 - loss: 0.0126 - val_Recall: 0.8597 - val_loss: 0.0643
Epoch 48/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9353 - loss: 0.0139 - val_Recall: 0.8597 - val_loss: 0.0626
Epoch 49/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9501 - loss: 0.0115 - val_Recall: 0.8741 - val_loss: 0.0647
Epoch 50/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9421 - loss: 0.0144 - val_Recall: 0.8741 - val_loss: 0.0631

print("Time taken in seconds ",end-start)

Time taken in seconds  80.87897825241089

plot(history,'loss')

plot(history,'Recall')

model_4_train_perf = model_performance_classification(model_4,X_train,y_train)
model_4_train_perf

469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 1ms/step

model_4_val_perf = model_performance_classification(model_4,X_val,y_val)
model_4_val_perf

157/157 ━━━━━━━━━━━━━━━━━━━━ 0s 1ms/step

y_train_pred_4 = model_4.predict(X_train)
y_val_pred_4 = model_4.predict(X_val)

469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 1ms/step
157/157 ━━━━━━━━━━━━━━━━━━━━ 0s 2ms/step

print("Classification Report - Train data Model_4", end="\n\n")
cr_train_model_4 = classification_report(y_train,y_train_pred_4 > 0.5)
print(cr_train_model_4)

Classification Report - Train data Model_4

              precision    recall  f1-score   support

           0       1.00      1.00      1.00     14168
           1       0.99      0.95      0.97       832

    accuracy                           1.00     15000
   macro avg       0.99      0.98      0.98     15000
weighted avg       1.00      1.00      1.00     15000

print("Classification Report - Validation data Model_4", end="\n\n")
cr_val_model_4 = classification_report(y_val,y_val_pred_4 > 0.5)
print(cr_val_model_4)

Classification Report - Validation data Model_4

              precision    recall  f1-score   support

           0       0.99      1.00      0.99      4722
           1       0.95      0.87      0.91       278

    accuracy                           0.99      5000
   macro avg       0.97      0.94      0.95      5000
weighted avg       0.99      0.99      0.99      5000

# clears the current Keras session, resetting all layers and models previously created, freeing up memory and resources.
tf.keras.backend.clear_session()

#Initializing the neural network
from tensorflow.keras.layers import Dropout
model_5 = Sequential()
model_5.add(Dense(64 ,activation="relu",input_dim=X_train.shape[1]))
model_5.add(Dropout(0.3))
model_5.add(Dense(32 ,activation="relu"))
model_5.add(Dense(16 , activation = "tanh"))
model_5.add(Dense(1 ,activation="sigmoid"))

model_5.summary()

Model: "sequential"

┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Layer (type)                    ┃ Output Shape           ┃       Param # ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ dense (Dense)                   │ (None, 64)             │         2,624 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dropout (Dropout)               │ (None, 64)             │             0 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_1 (Dense)                 │ (None, 32)             │         2,080 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_2 (Dense)                 │ (None, 16)             │           528 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_3 (Dense)                 │ (None, 1)              │            17 │
└─────────────────────────────────┴────────────────────────┴───────────────┘

 Total params: 5,249 (20.50 KB)

 Trainable params: 5,249 (20.50 KB)

 Non-trainable params: 0 (0.00 B)

# Set training parameters
epochs = 50
batch_size = 32

optimizer = tf.keras.optimizers.Adam()
model_5.compile(loss='binary_crossentropy', optimizer=optimizer, metrics = ['Recall'])

start = time.time()
history = model_5.fit(X_train, y_train, validation_data=(X_val,y_val) , batch_size=batch_size, epochs=epochs)
end=time.time()

Epoch 1/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 3s 3ms/step - Recall: 0.2898 - loss: 0.1580 - val_Recall: 0.7410 - val_loss: 0.0650
Epoch 2/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.7057 - loss: 0.0707 - val_Recall: 0.8273 - val_loss: 0.0543
Epoch 3/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8044 - loss: 0.0549 - val_Recall: 0.8381 - val_loss: 0.0501
Epoch 4/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.8183 - loss: 0.0558 - val_Recall: 0.8453 - val_loss: 0.0482
Epoch 5/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 3s 4ms/step - Recall: 0.8381 - loss: 0.0442 - val_Recall: 0.8525 - val_loss: 0.0475
Epoch 6/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8680 - loss: 0.0402 - val_Recall: 0.8633 - val_loss: 0.0463
Epoch 7/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8518 - loss: 0.0466 - val_Recall: 0.8417 - val_loss: 0.0460
Epoch 8/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8772 - loss: 0.0374 - val_Recall: 0.8489 - val_loss: 0.0465
Epoch 9/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8403 - loss: 0.0479 - val_Recall: 0.8597 - val_loss: 0.0457
Epoch 10/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8861 - loss: 0.0374 - val_Recall: 0.8633 - val_loss: 0.0449
Epoch 11/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8763 - loss: 0.0388 - val_Recall: 0.8669 - val_loss: 0.0449
Epoch 12/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8816 - loss: 0.0363 - val_Recall: 0.8705 - val_loss: 0.0435
Epoch 13/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 5ms/step - Recall: 0.8785 - loss: 0.0396 - val_Recall: 0.8669 - val_loss: 0.0456
Epoch 14/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.8825 - loss: 0.0410 - val_Recall: 0.8633 - val_loss: 0.0447
Epoch 15/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9028 - loss: 0.0333 - val_Recall: 0.8705 - val_loss: 0.0441
Epoch 16/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8921 - loss: 0.0355 - val_Recall: 0.8633 - val_loss: 0.0446
Epoch 17/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8838 - loss: 0.0374 - val_Recall: 0.8669 - val_loss: 0.0449
Epoch 18/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8936 - loss: 0.0355 - val_Recall: 0.8813 - val_loss: 0.0434
Epoch 19/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8945 - loss: 0.0374 - val_Recall: 0.8741 - val_loss: 0.0431
Epoch 20/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8873 - loss: 0.0350 - val_Recall: 0.8813 - val_loss: 0.0416
Epoch 21/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.8738 - loss: 0.0376 - val_Recall: 0.8813 - val_loss: 0.0431
Epoch 22/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 5ms/step - Recall: 0.8863 - loss: 0.0379 - val_Recall: 0.8741 - val_loss: 0.0443
Epoch 23/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - Recall: 0.8960 - loss: 0.0346 - val_Recall: 0.8777 - val_loss: 0.0434
Epoch 24/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8891 - loss: 0.0369 - val_Recall: 0.8813 - val_loss: 0.0426
Epoch 25/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8907 - loss: 0.0375 - val_Recall: 0.8813 - val_loss: 0.0422
Epoch 26/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8867 - loss: 0.0357 - val_Recall: 0.8741 - val_loss: 0.0434
Epoch 27/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8791 - loss: 0.0383 - val_Recall: 0.8777 - val_loss: 0.0438
Epoch 28/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 3s 3ms/step - Recall: 0.8923 - loss: 0.0353 - val_Recall: 0.8777 - val_loss: 0.0440
Epoch 29/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.8854 - loss: 0.0354 - val_Recall: 0.8813 - val_loss: 0.0437
Epoch 30/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 5ms/step - Recall: 0.8953 - loss: 0.0335 - val_Recall: 0.8813 - val_loss: 0.0434
Epoch 31/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 3ms/step - Recall: 0.8970 - loss: 0.0321 - val_Recall: 0.8849 - val_loss: 0.0433
Epoch 32/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8879 - loss: 0.0356 - val_Recall: 0.8813 - val_loss: 0.0425
Epoch 33/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9124 - loss: 0.0290 - val_Recall: 0.8849 - val_loss: 0.0441
Epoch 34/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8969 - loss: 0.0361 - val_Recall: 0.8813 - val_loss: 0.0433
Epoch 35/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8942 - loss: 0.0331 - val_Recall: 0.8705 - val_loss: 0.0439
Epoch 36/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8946 - loss: 0.0338 - val_Recall: 0.8741 - val_loss: 0.0417
Epoch 37/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8935 - loss: 0.0345 - val_Recall: 0.8777 - val_loss: 0.0415
Epoch 38/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 5ms/step - Recall: 0.8986 - loss: 0.0345 - val_Recall: 0.8849 - val_loss: 0.0416
Epoch 39/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.8953 - loss: 0.0358 - val_Recall: 0.8777 - val_loss: 0.0414
Epoch 40/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8993 - loss: 0.0326 - val_Recall: 0.8885 - val_loss: 0.0410
Epoch 41/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9001 - loss: 0.0319 - val_Recall: 0.8885 - val_loss: 0.0424
Epoch 42/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8893 - loss: 0.0355 - val_Recall: 0.8849 - val_loss: 0.0420
Epoch 43/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9028 - loss: 0.0326 - val_Recall: 0.8849 - val_loss: 0.0432
Epoch 44/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9213 - loss: 0.0284 - val_Recall: 0.8849 - val_loss: 0.0436
Epoch 45/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8925 - loss: 0.0347 - val_Recall: 0.8849 - val_loss: 0.0422
Epoch 46/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.8933 - loss: 0.0350 - val_Recall: 0.8849 - val_loss: 0.0416
Epoch 47/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 5ms/step - Recall: 0.9192 - loss: 0.0270 - val_Recall: 0.8849 - val_loss: 0.0413
Epoch 48/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 2s 4ms/step - Recall: 0.9150 - loss: 0.0316 - val_Recall: 0.8849 - val_loss: 0.0402
Epoch 49/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.9127 - loss: 0.0315 - val_Recall: 0.8813 - val_loss: 0.0404
Epoch 50/50
469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 3ms/step - Recall: 0.8956 - loss: 0.0316 - val_Recall: 0.8813 - val_loss: 0.0406

print("Time taken in seconds ",end-start)

Time taken in seconds  83.61508584022522

plot(history,'loss')

plot(history,'Recall')

model_5_train_perf = model_performance_classification(model_5,X_train,y_train)
model_5_train_perf

469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 1ms/step

model_5_val_perf = model_performance_classification(model_5,X_val,y_val)
model_5_val_perf

157/157 ━━━━━━━━━━━━━━━━━━━━ 0s 1ms/step

y_train_pred_5 = model_5.predict(X_train)
y_val_pred_5 = model_5.predict(X_val)

469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 1ms/step
157/157 ━━━━━━━━━━━━━━━━━━━━ 0s 1ms/step

print("Classification Report - Train data Model_2", end="\n\n")
cr_train_model_5 = classification_report(y_train,y_train_pred_5 > 0.5)
print(cr_train_model_5)

Classification Report - Train data Model_2

              precision    recall  f1-score   support

           0       1.00      1.00      1.00     14168
           1       1.00      0.92      0.95       832

    accuracy                           1.00     15000
   macro avg       1.00      0.96      0.98     15000
weighted avg       1.00      1.00      1.00     15000

print("Classification Report - Validation data Model_2", end="\n\n")
cr_val_model_5 = classification_report(y_val,y_val_pred_5 > 0.5)
print(cr_val_model_5)

Classification Report - Validation data Model_2

              precision    recall  f1-score   support

           0       0.99      1.00      1.00      4722
           1       0.98      0.88      0.93       278

    accuracy                           0.99      5000
   macro avg       0.99      0.94      0.96      5000
weighted avg       0.99      0.99      0.99      5000

# clears the current Keras session, resetting all layers and models previously created, freeing up memory and resources.
tf.keras.backend.clear_session()

model_6 = Sequential()
model_6.add(Dense(32,activation="relu",input_dim=X_train.shape[1]))
model_6.add(Dropout(0.5))
model_6.add(Dense(16 ,activation="relu"))
model_6.add(Dense(8 , activation = "relu"))
model_6.add(Dense(1 ,activation="sigmoid"))

model_6.summary()

Model: "sequential"

┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Layer (type)                    ┃ Output Shape           ┃       Param # ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ dense (Dense)                   │ (None, 32)             │         1,312 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dropout (Dropout)               │ (None, 32)             │             0 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_1 (Dense)                 │ (None, 16)             │           528 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_2 (Dense)                 │ (None, 8)              │           136 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_3 (Dense)                 │ (None, 1)              │             9 │
└─────────────────────────────────┴────────────────────────┴───────────────┘

 Total params: 1,985 (7.75 KB)

 Trainable params: 1,985 (7.75 KB)

 Non-trainable params: 0 (0.00 B)

# Set training parameters
epochs = 50
batch_size = 32

optimizer = tf.keras.optimizers.SGD(learning_rate=0.01, momentum=0.9)
model_6.compile(loss='binary_crossentropy', optimizer=optimizer, metrics = ['Recall'])

start = time.time()
history = model_6.fit(X_train, y_train, validation_data=(X_val,y_val) , batch_size=batch_size, epochs=epochs,class_weight=cw_dict, verbose=0 )
end=time.time()

print("Time taken in seconds ",end-start)

Time taken in seconds  72.53119802474976

plot(history,'loss')

plot(history,'Recall')

model_6_train_perf = model_performance_classification(model_6,X_train,y_train)
model_6_train_perf

469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step

model_6_val_perf = model_performance_classification(model_6,X_val,y_val)
model_6_val_perf

157/157 ━━━━━━━━━━━━━━━━━━━━ 0s 1ms/step

y_train_pred_6 = model_6.predict(X_train)
y_val_pred_6 = model_6.predict(X_val)

469/469 ━━━━━━━━━━━━━━━━━━━━ 1s 1ms/step
157/157 ━━━━━━━━━━━━━━━━━━━━ 0s 1ms/step

print("Classification Report - Train data Model_3", end="\n\n")
cr_train_model_6 = classification_report(y_train,y_train_pred_6 > 0.5)
print(cr_train_model_6)

Classification Report - Train data Model_3

              precision    recall  f1-score   support

           0       0.98      0.98      0.98     14168
           1       0.65      0.61      0.63       832

    accuracy                           0.96     15000
   macro avg       0.82      0.79      0.80     15000
weighted avg       0.96      0.96      0.96     15000

print("Classification Report - Validation data Model_3", end="\n\n")
cr_val_model_6 = classification_report(y_val,y_val_pred_6 > 0.5)
print(cr_val_model_6)

Classification Report - Validation data Model_3

              precision    recall  f1-score   support

           0       0.98      0.98      0.98      4722
           1       0.67      0.62      0.64       278

    accuracy                           0.96      5000
   macro avg       0.82      0.80      0.81      5000
weighted avg       0.96      0.96      0.96      5000

# training performance comparison

models_train_comp_df = pd.concat(
    [
        model_0_train_perf.T,
        model_1_train_perf.T,
        model_2_train_perf.T,
        model_3_train_perf.T,
        model_4_train_perf.T,
        model_5_train_perf.T,
        model_6_train_perf.T

    ],
    axis=1,
)
models_train_comp_df.columns = [
    "Model 0",
    "Model 1",
    "Model 2",
    "Model 3",
    "Model 4",
    "Model 5",
    "Model 6"
]
print("Training set performance comparison:")
models_train_comp_df

Training set performance comparison:

# Validation performance comparison

models_val_comp_df = pd.concat(
    [
        model_0_val_perf.T,
        model_1_val_perf.T,
        model_2_val_perf.T,
        model_3_val_perf.T,
        model_4_val_perf.T,
        model_5_val_perf.T,
        model_6_val_perf.T

    ],
    axis=1,
)
models_val_comp_df.columns = [
    "Model 0",
    "Model 1",
    "Model 2",
    "Model 3",
    "Model 4",
    "Model 5",
    "Model 6"
]
print("Validation set performance comparison:")
models_val_comp_df

Validation set performance comparison:

best_model = model_5

# Test set performance for the best model
best_model_test_perf = model_performance_classification(best_model,X_test,y_test)
best_model_test_perf

157/157 ━━━━━━━━━━━━━━━━━━━━ 0s 3ms/step

y_test_pred_best = best_model.predict(X_test)

cr_test_best_model = classification_report(y_test, y_test_pred_best>0.5) # Check the classification report of best model on test data.
print(cr_test_best_model)

157/157 ━━━━━━━━━━━━━━━━━━━━ 0s 1ms/step
              precision    recall  f1-score   support

           0       0.99      1.00      1.00      4718
           1       0.98      0.87      0.92       282

    accuracy                           0.99      5000
   macro avg       0.99      0.94      0.96      5000
weighted avg       0.99      0.99      0.99      5000

	V1	V2	V3	V4	V5	V6	V7	V8	V9	V10	V11	V12	V13	V14	V15	V16	V17	V18	V19	V20	V21	V22	V23	V24	V25	V26	V27	V28	V29	V30	V31	V32	V33	V34	V35	V36	V37	V38	V39	V40
0	-4.4646	-4.6791	3.1015	0.5061	-0.2211	-2.0325	-2.9109	0.0507	-1.5224	3.7619	-5.7147	0.7359	0.9813	1.4179	-3.3758	-3.0473	0.3062	2.9141	2.2700	4.3949	-2.3883	0.6464	-1.1905	3.1330	0.6653	-2.5108	-0.0367	0.7262	-3.9822	-1.0726	1.6671	3.0597	-1.6904	2.8463	2.2352	6.6675	0.4438	-2.3692	2.9506	-3.4803
1	3.3659	3.6534	0.9097	-1.3675	0.3320	2.3589	0.7326	-4.3321	0.5657	-0.1011	1.9145	-0.9515	-1.2553	-2.7065	0.1932	-4.7694	-2.2053	0.9077	0.7569	-5.8337	-3.0651	1.5966	-1.7573	1.7664	-0.2671	3.6250	1.5003	-0.5857	0.7830	-0.2012	0.0249	-1.7955	3.0328	-2.4675	1.8946	-2.2978	-1.7310	5.9088	-0.3863	0.6162
2	-3.8318	-5.8244	0.6340	-2.4188	-1.7738	1.0168	-2.0989	-3.1732	-2.0819	5.3926	-0.7707	1.1067	1.1443	0.9433	-3.1638	-4.2478	-4.0389	3.6885	3.3112	1.0590	-2.1430	1.6501	-1.6606	1.6799	-0.4508	-4.5507	3.7388	1.1344	-2.0335	0.8408	-1.6004	-0.2571	0.8035	4.0862	2.2921	5.3608	0.3520	2.9400	3.8392	-4.3094
3	1.6181	1.8883	7.0461	-1.1473	0.0831	-1.5298	0.2073	-2.4936	0.3449	2.1186	-3.0530	0.4597	2.7045	-0.6361	-0.4537	-3.1740	-3.4043	-1.2815	1.5821	-1.9518	-3.5166	-1.2060	-5.6279	-1.8177	2.1241	5.2946	4.7481	-2.3085	-3.9630	-6.0287	4.9488	-3.5844	-2.5775	1.3638	0.6227	5.5501	-1.5268	0.1389	3.1014	-1.2774
4	-0.1114	3.8725	-3.7584	-2.9829	3.7927	0.5450	0.2054	4.8490	-1.8549	-6.2200	1.9983	4.7238	0.7091	-1.9894	-2.6327	4.1844	2.2454	3.7345	-6.3128	-5.3799	-0.8867	2.0617	9.4456	4.4900	-3.9451	4.5821	-8.7804	-3.3830	5.1065	6.7875	2.0442	8.2659	6.6292	-10.0687	1.2230	-3.2298	1.6869	-2.1639	-3.6446	6.5103

	V1	V2	V3	V4	V5	V6	V7	V8	V9	V10	V11	V12	V13	V14	V15	V16	V17	V18	V19	V20	V21	V22	V23	V24	V25	V26	V27	V28	V29	V30	V31	V32	V33	V34	V35	V36	V37	V38	V39	V40	Target
19995	-2.0713	-1.0883	-0.7962	-3.0117	-2.2875	2.8073	0.4814	0.1052	-0.5866	-2.8994	8.8684	1.7172	1.3578	-1.7771	0.7098	4.9449	-3.1005	-1.1992	-1.0846	-0.3650	3.1312	-3.9481	-3.5785	-8.1391	-1.9369	-1.3277	-0.4027	-1.7348	9.9965	6.9554	-3.9385	-8.2740	5.7450	0.5890	-0.6500	-3.0432	2.2165	0.6087	0.1782	2.9278	1
19996	2.8903	2.4831	5.6439	0.9371	-1.3809	0.4121	-1.5934	-5.7625	2.1501	0.2723	-2.0948	-1.5258	0.0716	-3.5401	-2.7620	-10.6322	-0.4952	1.7201	3.8716	-1.2096	-8.2221	2.1209	-5.4918	1.4523	1.4500	3.6847	1.0768	-0.3842	-0.8386	-0.7483	-1.0886	-4.1591	1.1815	-0.7424	5.3690	-0.6930	-1.6690	3.6600	0.8199	-1.9873	0
19997	-3.8970	-3.9424	-0.3514	-2.4175	1.1075	-1.5276	-3.5199	2.0548	-0.2340	-0.3577	-3.7820	2.1800	6.1118	1.9847	-8.3300	-1.6392	-0.9150	5.6723	-3.9242	2.1332	-4.5020	2.7772	5.7279	1.6198	-1.6997	-0.0419	-2.9231	-2.7602	-2.2538	2.5520	0.9819	7.1122	1.4761	-3.9537	1.8556	5.0292	2.0826	-6.4093	1.4771	-0.8741	0
19998	-3.1873	-10.0517	5.6960	-4.3701	-5.3548	-1.8730	-3.9472	0.6794	-2.3893	5.4568	1.5830	3.5715	9.2266	2.5536	-7.0391	-0.9936	-9.6649	1.1552	3.8769	3.5236	-7.0153	-0.1320	-3.4462	-4.8014	-0.8757	-3.8119	5.4221	-3.7323	0.6088	5.2565	1.9148	0.4028	3.1637	3.7521	8.5299	8.4506	0.2040	-7.1299	4.2494	-6.1123	0
19999	-2.6869	1.9612	6.1371	2.6001	2.6572	-4.2909	-2.3443	0.9740	-1.0275	0.4974	-9.5891	3.1766	1.0545	-1.4159	-4.6686	-5.4054	3.7198	2.8929	2.3286	1.4577	-6.4285	1.8182	0.8059	7.7860	0.3309	5.2574	-4.8674	-0.8189	-5.6674	-2.8610	4.6743	6.6208	-1.9888	-1.3489	3.9518	5.4497	-0.4554	-2.2021	1.6782	-1.9744	0

	V1	V2	V3	V4	V5	V6	V7	V8	V9	V10	V11	V12	V13	V14	V15	V16	V17	V18	V19	V20	V21	V22	V23	V24	V25	V26	V27	V28	V29	V30	V31	V32	V33	V34	V35	V36	V37	V38	V39	V40
0	-0.6135	-3.8196	2.2023	1.3004	-1.1849	-4.4960	-1.8358	4.7230	1.2061	-0.3419	-5.1229	1.0170	4.8185	3.2690	-2.9843	1.3874	2.0320	-0.5116	-1.0231	7.3387	-2.2422	0.1555	2.0538	-2.7723	1.8514	-1.7887	-0.2773	-1.2551	-3.8329	-1.5045	1.5868	2.2912	-5.4114	0.8701	0.5745	4.1572	1.4281	-10.5113	0.4547	-1.4484
1	0.3896	-0.5123	0.5271	-2.5768	-1.0168	2.2351	-0.4413	-4.4057	-0.3329	1.9668	1.7965	0.4105	0.6383	-1.3896	-1.8834	-5.0179	-3.8272	2.4181	1.7623	-3.2423	-3.1930	1.8575	-1.7080	0.6334	-0.5879	0.0837	3.0139	-0.1823	0.2239	0.8652	-1.7822	-2.4749	2.4936	0.3152	2.0593	0.6839	-0.4855	5.1284	1.7207	-1.4882
2	-0.8749	-0.6406	4.0842	-1.5905	0.5259	-1.9576	-0.6954	1.3473	-1.7323	0.4665	-4.9282	3.5651	-0.4493	-0.6562	-0.1665	-1.6302	2.2919	2.3965	0.6013	1.7935	-2.1202	0.4820	-0.8407	1.7902	1.8744	0.3639	-0.1691	-0.4838	-2.1190	-2.1566	2.9073	-1.3189	-2.9975	0.4597	0.6198	5.6315	1.3235	-1.7522	1.8083	1.6757
3	0.2384	1.4586	4.0145	2.5345	1.1970	-3.1173	-0.9240	0.2695	1.3224	0.7023	-5.5783	-0.8507	2.5905	0.7674	-2.3908	-2.3420	0.5719	-0.9338	0.5087	1.2107	-3.2595	0.1046	-0.6589	1.4981	1.1003	4.1430	-0.2484	-1.1365	-5.3558	-4.5459	3.8087	3.5179	-3.0741	-0.2842	0.9546	3.0293	-1.3672	-3.4121	0.9060	-2.4509
4	5.8282	2.7683	-1.2345	2.8093	-1.6416	-1.4067	0.5686	0.9650	1.9184	-2.7749	-0.5300	1.3745	-0.6509	-1.6795	-0.3792	-4.4431	3.8939	-0.6076	2.9449	0.3672	-5.7891	4.5975	4.4503	3.2249	0.3967	0.2478	-2.3620	1.0794	-0.4731	2.2428	-3.5914	1.7738	-1.5016	-2.2267	4.7768	-6.5597	-0.8056	-0.2760	-3.8582	-0.5377

	V1	V2	V3	V4	V5	V6	V7	V8	V9	V10	V11	V12	V13	V14	V15	V16	V17	V18	V19	V20	V21	V22	V23	V24	V25	V26	V27	V28	V29	V30	V31	V32	V33	V34	V35	V36	V37	V38	V39	V40
4995	-5.1205	1.6348	1.2513	4.0359	3.2912	-2.9322	-1.3287	1.7541	-2.9846	1.2486	-6.8777	3.7152	-2.5118	-1.3950	-2.5541	-2.1974	4.7717	2.4029	3.7918	0.4868	-2.0281	1.7779	3.6680	11.3749	-1.9771	2.2516	-7.3185	1.9070	-3.7337	-0.0125	2.1205	9.9791	0.0634	0.2173	3.0364	2.1093	-0.5574	1.9387	0.5127	-2.6942
4996	-5.1725	1.1717	1.5791	1.2199	2.5296	-0.6686	-2.6183	-2.0005	0.6338	-0.5789	-3.6712	0.4602	3.3206	-1.0750	-7.1125	-4.3559	-0.0011	3.6984	-0.8464	-0.2223	-3.6450	0.7360	0.9257	3.2777	-2.2768	4.4576	-4.5429	-1.3480	-1.7794	0.3522	-0.2143	4.4239	2.6038	-2.1522	0.9174	2.1566	0.4670	0.4701	2.1968	-2.3765
4997	-1.1141	-0.4036	-1.7649	-5.8795	3.5716	3.7108	-2.4830	-0.3076	-0.9219	-2.9991	-0.1117	-1.9769	-1.6230	-0.9455	-2.7351	-0.8130	0.6098	8.1488	-9.1992	-3.8725	-0.2957	1.4684	2.8843	2.7921	-1.1357	1.1979	-4.3417	-2.8694	4.1241	4.1967	3.4707	3.7918	7.4815	-10.0614	-0.3872	1.8485	1.8182	-1.2456	-1.2609	7.4747
4998	-1.7032	0.6146	6.2205	-0.1041	0.9559	-3.2787	-1.6339	-0.1039	1.3882	-1.0656	-7.9698	2.2621	3.1340	-0.4858	-3.4983	-4.5617	3.1358	2.5364	-0.7922	4.3984	-4.0730	-0.0376	-2.3713	-1.5420	2.9079	3.2145	-0.1687	-1.5413	-4.7244	-5.5250	1.6680	-4.1004	-5.9493	0.5504	-1.5736	6.8239	2.1393	-4.0362	3.4361	0.5792
4999	-0.6037	0.9595	-0.7210	8.2296	-1.8156	-2.2755	-2.5745	-1.0415	4.1296	-2.7313	-3.2924	-1.6738	0.4645	-1.6459	-5.2634	-7.9876	6.4805	0.2263	4.9633	6.7520	-6.3058	3.2705	1.8972	3.2708	-0.6371	-0.9250	-6.7589	2.9902	-0.8138	3.4990	-8.4347	2.3698	-1.0624	0.7908	4.9520	-7.4408	-0.0695	-0.9181	-2.2912	-5.3629

	count	mean	std	min	25%	50%	75%	max
V1	19982.0000	-0.2720	3.4416	-11.8765	-2.7371	-0.7479	1.8401	15.4930
V2	19982.0000	0.4404	3.1508	-12.3200	-1.6407	0.4715	2.5440	13.0893
V3	20000.0000	2.4847	3.3890	-10.7081	0.2069	2.2558	4.5662	17.0909
V4	20000.0000	-0.0832	3.4316	-15.0821	-2.3477	-0.1352	2.1306	13.2364
V5	20000.0000	-0.0538	2.1048	-8.6034	-1.5356	-0.1020	1.3405	8.1338
V6	20000.0000	-0.9954	2.0410	-10.2271	-2.3472	-1.0005	0.3803	6.9758
V7	20000.0000	-0.8793	1.7616	-7.9497	-2.0309	-0.9172	0.2237	8.0061
V8	20000.0000	-0.5482	3.2958	-15.6576	-2.6427	-0.3891	1.7230	11.6795
V9	20000.0000	-0.0168	2.1606	-8.5963	-1.4950	-0.0676	1.4092	8.1376
V10	20000.0000	-0.0130	2.1932	-9.8540	-1.4112	0.1010	1.4770	8.1085
V11	20000.0000	-1.8954	3.1243	-14.8321	-3.9224	-1.9212	0.1189	11.8264
V12	20000.0000	1.6048	2.9305	-12.9480	-0.3965	1.5078	3.5715	15.0807
V13	20000.0000	1.5805	2.8747	-13.2282	-0.2235	1.6372	3.4599	15.4196
V14	20000.0000	-0.9506	1.7897	-7.7386	-2.1707	-0.9572	0.2707	5.6707
V15	20000.0000	-2.4150	3.3550	-16.4166	-4.4153	-2.3826	-0.3591	12.2465
V16	20000.0000	-2.9252	4.2217	-20.3742	-5.6342	-2.6827	-0.0950	13.5832
V17	20000.0000	-0.1343	3.3455	-14.0912	-2.2156	-0.0146	2.0688	16.7564
V18	20000.0000	1.1893	2.5923	-11.6440	-0.4039	0.8834	2.5718	13.1799
V19	20000.0000	1.1818	3.3969	-13.4918	-1.0502	1.2791	3.4933	13.2377
V20	20000.0000	0.0236	3.6695	-13.9227	-2.4330	0.0334	2.5124	16.0523
V21	20000.0000	-3.6113	3.5677	-17.9562	-5.9304	-3.5329	-1.2659	13.8405
V22	20000.0000	0.9518	1.6515	-10.1221	-0.1181	0.9747	2.0256	7.4099
V23	20000.0000	-0.3661	4.0319	-14.8661	-3.0988	-0.2621	2.4517	14.4587
V24	20000.0000	1.1344	3.9121	-16.3871	-1.4681	0.9690	3.5460	17.1633
V25	20000.0000	-0.0022	2.0167	-8.2283	-1.3652	0.0251	1.3971	8.2234
V26	20000.0000	1.8738	3.4351	-11.8343	-0.3379	1.9505	4.1300	16.8364
V27	20000.0000	-0.6124	4.3688	-14.9049	-3.6523	-0.8849	2.1892	17.5604
V28	20000.0000	-0.8832	1.9177	-9.2695	-2.1712	-0.8911	0.3759	6.5276
V29	20000.0000	-0.9856	2.6844	-12.5795	-2.7874	-1.1762	0.6298	10.7221
V30	20000.0000	-0.0155	3.0053	-14.7960	-1.8671	0.1843	2.0362	12.5058
V31	20000.0000	0.4868	3.4614	-13.7228	-1.8178	0.4903	2.7307	17.2551
V32	20000.0000	0.3038	5.5004	-19.8765	-3.4205	0.0521	3.7617	23.6332
V33	20000.0000	0.0498	3.5753	-16.8984	-2.2429	-0.0662	2.2551	16.6925
V34	20000.0000	-0.4627	3.1838	-17.9851	-2.1370	-0.2550	1.4369	14.3582
V35	20000.0000	2.2296	2.9371	-15.3498	0.3362	2.0986	4.0644	15.2911
V36	20000.0000	1.5148	3.8009	-14.8332	-0.9438	1.5665	3.9839	19.3296
V37	20000.0000	0.0113	1.7882	-5.4784	-1.2558	-0.1284	1.1755	7.4670
V38	20000.0000	-0.3440	3.9481	-17.3750	-2.9876	-0.3168	2.2794	15.2899
V39	20000.0000	0.8907	1.7531	-6.4389	-0.2723	0.9193	2.0575	7.7599
V40	20000.0000	-0.8756	3.0122	-11.0239	-2.9402	-0.9208	1.1199	10.6543
Target	20000.0000	0.0555	0.2290	0.0000	0.0000	0.0000	0.0000	1.0000

Problem Statement¶

Business Context¶

Objective¶

Data Description¶

Installing and Importing the necessary libraries¶

Loading the Data¶

Data Overview¶

Observations:¶

Observations on Target Distribution:¶

Observations:¶

Observations:¶

Exploratory Data Analysis¶

Univariate analysis¶

Obserations:¶

Bivariate Analysis¶

Observations:¶

Observations:¶

Data Preprocessing¶

Observations:¶

Model Building¶

Model Evaluation Criterion¶

Initial Model Building (Model 0)¶

Observations:¶

Model Performance Improvement¶

Model 1¶

Observations:¶

Model 2¶

Observations:¶

Model 3¶

Observations:¶

Model 4¶

Observations:¶

Model 5¶

Observations:¶

Model 6¶

Observations:¶

Model Performance Comparison and Final Model Selection¶

Best Model:¶

Actionable Insights and Recommendations¶

	count	mean	std	min	25%	50%	75%	max
V1	4995.0000	-0.2776	3.4663	-12.3817	-2.7437	-0.7648	1.8313	13.5044
V2	4994.0000	0.3979	3.1396	-10.7162	-1.6492	0.4274	2.4445	14.0791
V3	5000.0000	2.5518	3.3266	-9.2379	0.3149	2.2604	4.5870	15.3145
V4	5000.0000	-0.0489	3.4139	-14.6824	-2.2927	-0.1458	2.1665	12.1402
V5	5000.0000	-0.0801	2.1109	-7.7116	-1.6152	-0.1319	1.3412	7.6728
V6	5000.0000	-1.0421	2.0054	-8.9242	-2.3689	-1.0486	0.3076	5.0677
V7	5000.0000	-0.9079	1.7690	-8.1242	-2.0543	-0.9397	0.2122	7.6162
V8	5000.0000	-0.5746	3.3319	-12.2527	-2.6421	-0.3579	1.7129	10.4147
V9	5000.0000	0.0301	2.1741	-6.7855	-1.4557	-0.0799	1.4495	8.8507
V10	5000.0000	0.0185	2.1454	-8.1710	-1.3533	0.1663	1.5112	6.5987
V11	5000.0000	-2.0086	3.1122	-13.1518	-4.0504	-2.0431	0.0441	9.9564
V12	5000.0000	1.5764	2.9074	-8.1640	-0.4497	1.4883	3.5626	12.9836
V13	5000.0000	1.6225	2.8829	-11.5482	-0.1260	1.7186	3.4646	12.6200
V14	5000.0000	-0.9211	1.8035	-7.8139	-2.1110	-0.8960	0.2723	5.7341
V15	5000.0000	-2.4522	3.3870	-15.2858	-4.4791	-2.4171	-0.4329	11.6734
V16	5000.0000	-3.0185	4.2644	-20.9858	-5.6483	-2.7738	-0.1781	13.9758
V17	5000.0000	-0.1037	3.3365	-13.4183	-2.2277	0.0475	2.1119	19.7766
V18	5000.0000	1.1956	2.5864	-12.2140	-0.4089	0.8814	2.6040	13.6422
V19	5000.0000	1.2105	3.3847	-14.1696	-1.0264	1.2959	3.5263	12.4280
V20	5000.0000	0.1384	3.6572	-13.7196	-2.3255	0.1934	2.5396	13.8706
V21	5000.0000	-3.6644	3.5778	-16.3407	-5.9444	-3.6629	-1.3296	11.0469
V22	5000.0000	0.9620	1.6404	-6.7402	-0.0477	0.9860	2.0293	7.5053
V23	5000.0000	-0.4222	4.0567	-14.4223	-3.1627	-0.2792	2.4259	13.1809
V24	5000.0000	1.0888	3.9682	-12.3155	-1.6232	0.9128	3.5372	17.8060
V25	5000.0000	0.0612	2.0102	-6.7701	-1.2984	0.0767	1.4285	6.5569
V26	5000.0000	1.8473	3.4003	-11.4140	-0.2425	1.9170	4.1561	17.5282
V27	5000.0000	-0.5524	4.4029	-13.1770	-3.6626	-0.8720	2.2473	17.2902
V28	5000.0000	-0.8677	1.9262	-7.9334	-2.1598	-0.9307	0.4206	7.4157
V29	5000.0000	-1.0958	2.6555	-9.9878	-2.8614	-1.3405	0.5218	14.0395
V30	5000.0000	-0.1187	3.0233	-12.4384	-1.9967	0.1125	1.9465	10.3150
V31	5000.0000	0.4688	3.4463	-11.2633	-1.8224	0.4857	2.7790	12.5589
V32	5000.0000	0.2326	5.5856	-17.2442	-3.5563	-0.0767	3.7519	26.5394
V33	5000.0000	-0.0801	3.5386	-14.9038	-2.3481	-0.1597	2.0992	13.3235
V34	5000.0000	-0.3927	3.1661	-14.6997	-2.0096	-0.1717	1.4654	12.1463
V35	5000.0000	2.2112	2.9484	-12.2606	0.3218	2.1117	4.0316	13.4892
V36	5000.0000	1.5948	3.7750	-12.7356	-0.8661	1.7030	4.1044	17.1161
V37	5000.0000	0.0229	1.7853	-5.0791	-1.2405	-0.1104	1.2375	6.8099
V38	5000.0000	-0.4057	3.9689	-15.3345	-2.9845	-0.3812	2.2880	13.0650
V39	5000.0000	0.9388	1.7165	-5.4510	-0.2080	0.9592	2.1308	7.1822
V40	5000.0000	-0.9324	2.9782	-10.0762	-2.9866	-1.0028	1.0797	8.6985
Target	5000.0000	0.0564	0.2307	0.0000	0.0000	0.0000	0.0000	1.0000

Aspect	Finding	Remark
Model fit	Excellent	Balanced loss curves
Overfitting	None	Train ≈ Validation performance
Recall	~0.81–0.82	Could be optimized further
F1-score	~0.94	High overall model quality
Business alignment	Good	Needs minor recall optimization

Aspect	Model_0 (Simple NN)	Model_1 (Deeper NN)	Observation
Layers	7 → 1	128 → 7 → 1	Added complexity
Training Time	69.6s	48s	Slightly faster (larger batch size)
Accuracy	0.987	0.991	Improved
Recall	0.81–0.90	0.86–0.94	Strong improvement
Precision	0.97	0.98–0.99	Improved stability
Overfitting	None	None	Stable performance
Business Value	Good detection	Better failure capture	Fewer missed failures

Metric	Model_0	Model_1	Model_2
Architecture	Simple (7→1)	128→7→1	16→Dropout→128→64→1
Parameters	295	6,159	11,153
Training Time	69.6s	48s	77s
Accuracy (Val)	0.987	0.991	0.986
Recall (Val)	0.81	0.93	0.88
Precision (Val)	0.97	0.98	0.97
F1-score (Val)	0.88	0.95	0.92
Overfitting	None	None	None (slightly underfit)
Generalization	Good	Excellent	Good but conservative

Goal	Recommendation
Best-performing model	✅ Model_1 (128 → 7 → 1)
If avoiding overfitting is priority	Use Model_2 with Dropout=0.3 instead of 0.5

Metric	Model_0	Model_1	Model_2	Model_3 (Class Weighted)
Architecture	7→1	128→7→1	16→Dropout→128→64→1	16→Dropout→128→64→1
Params	295	6,159	11,153	11,153
Time (s)	69.6	48	77.3	74.9
Accuracy (Val)	0.987	0.991	0.986	0.965
Recall (Val)	0.81	0.93	0.77	0.88–0.89
Precision (Val)	0.97	0.98	0.97	0.81
F1-score (Val)	0.88	0.95	0.92	0.86
Overfitting	None	None	None	None
Handling Imbalance	❌	❌	❌	✅
Business Impact	Good	Excellent	Conservative	High sensitivity to failures

Business Priority	Recommended Model	Reason
Cost-efficient operation	✅ Model_3	High recall ensures early detection and preventive maintenance.
Balanced performance (recall + precision)	Model_1	Best overall metrics and generalization.

Goal	Recommended Model	Reason
Best overall model	✅ Model_4	Excellent recall (0.94), precision (0.97), and generalization.
High recall, more preventive action	Model_3	Slightly higher recall sensitivity (with class weights).
Balanced business trade-off	Model_4	Best cost-effective solution (few misses + minimal false alarms).

Objective	Best Model	Rationale
Overall balance (accuracy, recall, precision)	✅ Model_5	Highest F1-score (0.96), minimal overfitting, and efficient architecture.
High recall (detect all failures)	Model_4	Slightly higher recall but heavier model.
Operational efficiency	Model_5	Lightweight yet accurate → ideal for deployment on production systems.

Metric	Class 0 (No Failure)	Class 1 (Failure)	Overall
Precision	0.98	0.65	0.96
Recall	0.98	0.61	0.96
F1-Score	0.98	0.63	0.96
Accuracy	—	—	0.96

Metric	Score
Accuracy	0.9931
Recall	0.9432
Precision	0.9907
F1-score	0.9656

Metric	Score
Accuracy	0.9906
Recall	0.9273
Precision	0.9815
F1-score	0.9526

Metric	Score
Accuracy	0.9849
Recall	0.8715
Precision	0.9815
F1-score	0.9185

Metric	Score
Accuracy	0.9856
Recall	0.8857
Precision	0.9733
F1-score	0.9245

	Model 0	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Accuracy	0.9874	0.9931	0.9849	0.9685	0.9968	0.9952	0.9605
Recall	0.9034	0.9432	0.8715	0.9403	0.9757	0.9579	0.7947
Precision	0.9731	0.9907	0.9815	0.8251	0.9936	0.9962	0.8158
F1 Score	0.9351	0.9656	0.9185	0.8724	0.9844	0.9762	0.8048

	Model 0	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Accuracy	0.9878	0.9906	0.9856	0.9646	0.9904	0.9924	0.9616
Recall	0.9038	0.9273	0.8857	0.9254	0.9357	0.9401	0.8002
Precision	0.9773	0.9815	0.9733	0.8113	0.9709	0.9865	0.8222
F1 Score	0.9371	0.9526	0.9245	0.8578	0.9525	0.9620	0.8108

Metric	Model 0	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Accuracy	0.9878	0.9906	0.9856	0.9646	0.9904	0.9924	0.9616
Recall	0.9038	0.9273	0.8857	0.9254	0.9357	0.9401	0.8002
Precision	0.9773	0.9815	0.9733	0.8113	0.9709	0.9865	0.8222
F1 Score	0.9371	0.9526	0.9245	0.8578	0.9525	0.9620	0.8108

Model	Key Traits	Verdict
Model 3	High recall but low precision, unstable training	❌ Underfitting
Model 4	Very strong training but mild overfitting	⚠️ Slightly overfitted
Model 5	Best balance of recall, precision, F1, and stability	🏆 Final Choice
Model 6	Lowest recall, unstable loss	❌ Poor convergence

Metric	Value
Accuracy	0.9918
Recall	0.9356
Precision	0.9862
F1 Score	0.9594