ann

import pandas as pd
import seaborn as sns

df = pd.read_csv('./Churn_Modelling.csv')

df.shape

df.columns

df.head()

x = df[['CreditScore','Age','Tenure','Balance','NumOfProducts','HasCrCard',
        'IsActiveMember','EstimatedSalary']]

y = df['Exited']

x

sns.countplot(x = y);

y.value_counts()

from sklearn.preprocessing import StandardScaler

scaler = StandardScaler()

x_scaled = scaler.fit_transform(x)

x_scaled

from sklearn.model_selection import train_test_split

x_train, x_test, y_train, y_test = train_test_split(x_scaled,y,random_state =0, test_size=0.25)

x.shape

x_test.shape

x_train.shape

from sklearn.neural_network import MLPClassifier

ann = MLPClassifier(hidden_layer_sizes=(100,100,100), random_state=0,
                    max_iter=100, activation = 'relu')

ann.fit(x_train, y_train)

y_pred = ann.predict(x_test)

y_pred

from sklearn.metrics import ConfusionMatrixDisplay, classification_report
from sklearn.metrics import accuracy_score

y_test.value_counts()

ConfusionMatrixDisplay.from_predictions(y_test, y_pred)

before = accuracy_score(y_test, y_pred)
before

print(classification_report(y_test, y_pred))

from imblearn.over_sampling import RandomOverSampler
ros = RandomOverSampler(random_state=0)

x_res, y_res = ros.fit_resample(x, y)

y_res.value_counts()

scaler = StandardScaler()
x_scaled = scaler.fit_transform(x_res)
x_scaled

x_train, x_test, y_train, y_test = train_test_split(x_scaled,y_res,random_state =0, test_size=0.25)

x_res.shape

x_test.shape

x_train.shape

ann = MLPClassifier(hidden_layer_sizes=(100,100,100), random_state=0,
                    max_iter=100, activation = 'relu')

ann.fit(x_train, y_train)

y_pred = ann.predict(x_test)
y_pred

y_test.value_counts()

ConfusionMatrixDisplay.from_predictions(y_test, y_pred)

after = accuracy_score(y_test, y_pred)
after

print(classification_report(y_test, y_pred))

import matplotlib.pyplot as plt

accuracy_scores = [before, after]

labels = ['Before Optimization', 'After Optimization']

plt.figure(figsize=(8, 5))
plt.plot(labels, accuracy_scores, marker='o', linestyle='-', color='b')
plt.title('Comparison of ANN Accuracy Scores')
plt.xlabel('Optimization Status')
plt.ylabel('Accuracy Score')
plt.ylim(0.8, 0.9)
plt.grid(True)
plt.show()