Models_with_PCA.py

import pandas as pd
import numpy as np
from Data_preproccesing import x, y
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score
from sklearn.neighbors import KNeighborsClassifier
from sklearn.tree import DecisionTreeClassifier
from sklearn.linear_model import LogisticRegression
from sklearn import svm
from sklearn.ensemble import RandomForestClassifier
from sklearn.decomposition import PCA
from imblearn.over_sampling import RandomOverSampler

# Assuming final_df and target are already defined

# Oversampling
ros = RandomOverSampler(random_state=0)
X_resampled, y_resampled = ros.fit_resample(x, y)

# Train-test split
X_train, X_test, y_train, y_test = train_test_split(X_resampled, y_resampled, test_size=0.2, random_state=1)

# Standardization
scaler = StandardScaler()
X_train_scaled = scaler.fit_transform(X_train)
X_test_scaled = scaler.transform(X_test)

# PCA
pca = PCA(n_components=0.98)
X_train_pca = pca.fit_transform(X_train_scaled)
X_test_pca = pca.transform(X_test_scaled)

# Initialize an empty DataFrame
result = pd.DataFrame(columns=['Model', 'Train Accuracy', 'Test Accuracy'])

# KNN with PCA
knn_classifier = KNeighborsClassifier()
knn_classifier.fit(X_train_pca, y_train)
knn_train_accuracy = accuracy_score(y_train, knn_classifier.predict(X_train_pca))
knn_test_accuracy = accuracy_score(y_test, knn_classifier.predict(X_test_pca))
result = result._append({'Model': 'KNN with PCA', 'Train Accuracy': knn_train_accuracy, 'Test Accuracy': knn_test_accuracy}, ignore_index=True)

# Logistic Regression with PCA
logistic_classifier = LogisticRegression(C=20, random_state=None)
logistic_classifier.fit(X_train_pca, y_train)
logistic_train_accuracy = accuracy_score(y_train, logistic_classifier.predict(X_train_pca))
logistic_test_accuracy = accuracy_score(y_test, logistic_classifier.predict(X_test_pca))
result = result._append({'Model': 'Logistic Regression with PCA', 'Train Accuracy': logistic_train_accuracy, 'Test Accuracy': logistic_test_accuracy}, ignore_index=True)

# Decision Tree with PCA
dt_classifier = DecisionTreeClassifier(criterion='entropy', random_state=0)
dt_classifier.fit(X_train_pca, y_train)
dt_train_accuracy = accuracy_score(y_train, dt_classifier.predict(X_train_pca))
dt_test_accuracy = accuracy_score(y_test, dt_classifier.predict(X_test_pca))
result = result._append({'Model': 'Decision Tree with PCA', 'Train Accuracy': dt_train_accuracy, 'Test Accuracy': dt_test_accuracy}, ignore_index=True)

# Linear SVM with PCA
svm_classifier = svm.SVC(kernel='linear', random_state=0)
svm_classifier.fit(X_train_pca, y_train)
svm_train_accuracy = accuracy_score(y_train, svm_classifier.predict(X_train_pca))
svm_test_accuracy = accuracy_score(y_test, svm_classifier.predict(X_test_pca))
result = result._append({'Model': 'Linear SVM with PCA', 'Train Accuracy': svm_train_accuracy, 'Test Accuracy': svm_test_accuracy}, ignore_index=True)

# Random Forest with PCA
rf_classifier = RandomForestClassifier(n_estimators=100)
rf_classifier.fit(X_train_pca, y_train)
rf_train_accuracy = accuracy_score(y_train, rf_classifier.predict(X_train_pca))
rf_test_accuracy = accuracy_score(y_test, rf_classifier.predict(X_test_pca))
result = result._append({'Model': 'Random Forest with PCA', 'Train Accuracy': rf_train_accuracy, 'Test Accuracy': rf_test_accuracy}, ignore_index=True)

print(result)


# import matplotlib.pyplot as plt

# # Plot train and test accuracy for each model
# plt.figure(figsize=(10, 6))
# plt.bar(result['Model'], result['Train Accuracy'], color='b', alpha=0.5, label='Train Accuracy')
# plt.bar(result['Model'], result['Test Accuracy'], color='r', alpha=0.5, label='Test Accuracy')
# plt.xticks(rotation=45, ha='right')
# plt.ylabel('Accuracy')
# plt.title('Train and Test Accuracy for Different Models')
# plt.legend()
# plt.tight_layout()
# plt.show()