Abdelrahma12/heart

###Importing Liberaries

import numpy as np import pandas as pd import matplotlib.pyplot as plt import seaborn as sns from sklearn.model_selection import GridSearchCV from sklearn.linear_model import LogisticRegression from sklearn.ensemble import RandomForestClassifier from sklearn.neural_network import MLPClassifier from sklearn.neighbors import KNeighborsClassifier from xgboost import XGBClassifier from sklearn.svm import SVC from sklearn.metrics import accuracy_score, classification_report import warnings warnings.filterwarnings('ignore')

from google.colab import drive

Data Load

data = pd.read_csv(r'/content/heart_failure_clinical_records_dataset - Copy.csv')

Data Exploratory

Data Exploratory

data

data.head()

data.info()

data.duplicated().sum()

labels = ["40-45", "46-50", "51-55", "56-60", "61-65", "66-70", "71-75", "76-80", "81-95"] data['age_group'] = pd.cut(data['age'], bins=[40, 45, 50, 55, 60, 65, 70, 75, 80, 95], labels=labels)

data.isnull().sum()

Data Visualization

plt.figure(figsize=(10,6)) sns.countplot(data=data, x='age_group', hue='DEATH_EVENT', palette=["lightblue", "red"]) plt.title("Death Count by Age Group") plt.xlabel("Age Group") plt.ylabel("Patient Count") plt.legend(["Survived", "Died"]) plt.show()

corr_matrix = data.drop(columns=['age_group']).corr() plt.figure(figsize=(12, 10)) sns.heatmap(corr_matrix, annot=True, cmap='coolwarm', fmt=".2f") plt.title('Correlation Matrix of Heart Failure Clinical Records') plt.show()

death_counts = data['DEATH_EVENT'].value_counts() plt.figure(figsize=(6, 6)) plt.pie(death_counts, labels=['Not Died', 'Died'], autopct='%1.1f%%', startangle=90, colors=['skyblue', 'lightcoral']) plt.title('Distribution of DEATH_EVENT') plt.show()

Select a subset of numerical features that showed some correlation with DEATH_EVENT

selected_features = ['time', 'serum_creatinine', 'ejection_fraction', 'age', 'serum_sodium', 'DEATH_EVENT']

sns.pairplot(data[selected_features], hue='DEATH_EVENT', diag_kind='kde') plt.suptitle('Pairplot of Selected Numerical Features by DEATH_EVENT', y=1.02) plt.show()

Data Preprocessing

Data Split

data.drop(columns=['age_group'], inplace=True)

X = data.drop('DEATH_EVENT', axis=1) y = data['DEATH_EVENT'] from sklearn.model_selection import train_test_split X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42, stratify=y)

Feature Scaling

from sklearn.preprocessing import StandardScaler scaler = StandardScaler() continuous_features = ['age', 'creatinine_phosphokinase', 'ejection_fraction', 'platelets', 'serum_creatinine', 'serum_sodium', 'time'] X_train[continuous_features] = scaler.fit_transform(X_train[continuous_features]) X_test[continuous_features] = scaler.transform(X_test[continuous_features])

#Modeling

Logistic Regression

log_params = { 'penalty': ['l1', 'l2', 'elasticnet', 'none'], 'C': [0.01, 0.1, 1, 10, 100], 'solver': ['lbfgs', 'saga'], 'max_iter': [1000] }

log_grid = GridSearchCV(LogisticRegression(random_state=42), log_params, cv=5) log_grid.fit(X_train, y_train)

print(" Logistic Regression Best Params:", log_grid.best_params_)

####Evaluation

log_model = LogisticRegression( penalty='l2', C=0.1, solver='lbfgs', max_iter=1000, random_state=42 ) log_model.fit(X_train, y_train) y_pred_log = log_model.predict(X_test) print(" Logistic Regression") print(f"Accuracy: {accuracy_score(y_test, y_pred_log):.4f}") print(classification_report(y_test, y_pred_log))

Random Forest

rf_params = { 'n_estimators': [50, 100, 200], 'max_depth': [None, 5, 10], 'min_samples_split': [2, 5], 'min_samples_leaf': [1, 2] }

rf_grid = GridSearchCV(RandomForestClassifier(random_state=42), rf_params, cv=5) rf_grid.fit(X_train, y_train)

print(" Random Forest Best Params:", rf_grid.best_params_)

####Evaluation

rf_model = RandomForestClassifier( n_estimators=50, max_depth=5, min_samples_leaf=2, min_samples_split=5, random_state=42 ) rf_model.fit(X_train, y_train) y_pred_rf = rf_model.predict(X_test) print(" Random Forest") print(f"Accuracy: {accuracy_score(y_test, y_pred_rf):.4f}") print(classification_report(y_test, y_pred_rf))

SVM

svm_params = { 'kernel': ['linear', 'rbf'], 'C': [0.1, 1, 10], 'gamma': ['scale', 'auto'] }

svm_grid = GridSearchCV(SVC(probability=True, random_state=42), svm_params, cv=5) svm_grid.fit(X_train, y_train)

print(" SVM Best Params:", svm_grid.best_params_)

Evaluation

svm_model = SVC( C=0.1, gamma='scale', kernel='linear', probability=True, random_state=42 ) svm_model.fit(X_train, y_train) y_pred_svm = svm_model.predict(X_test) print("\n SVM") print(f"Accuracy: {accuracy_score(y_test, y_pred_svm):.4f}") print(classification_report(y_test, y_pred_svm))

MLP

mlp_params = { 'hidden_layer_sizes': [(64,), (64, 32), (128, 64)], 'activation': ['relu', 'tanh'], 'alpha': [0.0001, 0.001], 'learning_rate': ['constant', 'adaptive'] }

mlp_grid = GridSearchCV(MLPClassifier(max_iter=1000, random_state=42), mlp_params, cv=5) mlp_grid.fit(X_train, y_train)

print(" MLP Best Params:", mlp_grid.best_params_)

Evaluation

mlp_model = MLPClassifier( hidden_layer_sizes=(64, 32), activation='tanh', alpha=0.0001, learning_rate='constant', max_iter=1000, random_state=42 ) mlp_model.fit(X_train, y_train) y_pred_mlp = mlp_model.predict(X_test) print("\n MLP Neural Network") print(f"Accuracy: {accuracy_score(y_test, y_pred_mlp):.4f}") print(classification_report(y_test, y_pred_mlp))

XGBoost

xgb_params = { 'n_estimators': [50, 100, 200], 'max_depth': [3, 4, 5], 'learning_rate': [0.01, 0.1, 0.2] }

xgb_grid = GridSearchCV( XGBClassifier(use_label_encoder=False, eval_metric='logloss', random_state=42), xgb_params, cv=5 ) xgb_grid.fit(X_train, y_train)

print(" XGBoost Best Params:", xgb_grid.best_params_)

Evaluation

xgb_model = XGBClassifier( n_estimators=50, max_depth=4, learning_rate=0.2, use_label_encoder=False, eval_metric='logloss', random_state=42 ) xgb_model.fit(X_train, y_train) y_pred_xgb = xgb_model.predict(X_test) print("\n XGBoost") print(f"Accuracy: {accuracy_score(y_test, y_pred_xgb):.4f}") print(classification_report(y_test, y_pred_xgb))

KNN

knn_params = { 'n_neighbors': [3, 5, 7, 9], 'weights': ['uniform', 'distance'], 'metric': ['euclidean', 'manhattan'] }

knn_grid = GridSearchCV(KNeighborsClassifier(), knn_params, cv=5) knn_grid.fit(X_train, y_train)

print(" KNN Best Params:", knn_grid.best_params_)

Evaluation

knn_model = KNeighborsClassifier( n_neighbors=5, weights='uniform', metric='euclidean' ) knn_model.fit(X_train, y_train) y_pred_knn = knn_model.predict(X_test) print("\n KNN") print(f"Accuracy: {accuracy_score(y_test, y_pred_knn):.4f}") print(classification_report(y_test, y_pred_knn))

Models Accuracies

models = [ 'Random Forest', 'SVM', 'MLP', 'XGBoost', 'KNN', 'Logistic Regression' ] accuracies = [ 0.85, 0.8333, 0.6833, 0.8333, 0.7167, 0.8333 ]

plt.figure(figsize=(10, 6)) plt.bar(models, accuracies, color=['blue', 'green', 'purple', 'orange', 'red', 'cyan']) plt.ylim(0, 1) plt.ylabel('Accuracy') plt.title('Model Accuracy Comparison')

plt.xticks(rotation=30) plt.show()

import gradio as gr from sklearn.preprocessing import StandardScaler import joblib

joblib.dump(rf_model, "heart_model.pkl") joblib.dump(scaler, "scaler.pkl") print("Model and scaler saved successfully")

model = joblib.load("heart_model.pkl") scaler = joblib.load("scaler.pkl")

def predict_heart_risk(age, cpk, ef, platelets, sc, ss, time, anaemia, diabetes, high_bp, sex, smoking): data = pd.DataFrame([[ age, anaemia, cpk, diabetes, ef, high_bp, platelets, sc, ss, sex, smoking, time ]], columns=[ 'age', 'anaemia', 'creatinine_phosphokinase', 'diabetes', 'ejection_fraction', 'high_blood_pressure', 'platelets', 'serum_creatinine', 'serum_sodium', 'sex', 'smoking', 'time' ])

continuous_features = ['age', 'creatinine_phosphokinase', 'ejection_fraction','platelets', 'serum_creatinine', 'serum_sodium', 'time']
data[continuous_features] = scaler.transform(data[continuous_features])

prediction = model.predict(data)[0]
return " At Risk" if prediction == 1 else " Not At Risk"

inputs = [ gr.Number(label="Age"), gr.Number(label="Creatinine Phosphokinase, Range [0,100000]"), gr.Number(label="Ejection Fraction, Range [5,85] "), gr.Number(label="Platelets, Range [5000,2000000]"), gr.Number(label="Serum Creatinine, Range [0.1,60]"), gr.Number(label="Serum Sodium, Range [95,255]"), gr.Number(label="Follow-up Time (days)"), gr.Radio([0, 1], label="Anaemia (0=No, 1=Yes)"), gr.Radio([0, 1], label="Diabetes (0=No, 1=Yes)"), gr.Radio([0, 1], label="High Blood Pressure (0=No, 1=Yes)"), gr.Radio([0, 1], label="Sex (0=Female, 1=Male)"), gr.Radio([0, 1], label="Smoking (0=No, 1=Yes)") ]

gr.Interface( fn=predict_heart_risk, inputs=inputs, outputs="text", title=" Heart Failure Risk Predictor", description="Enter patient data to predict if they are at risk of heart failure.", allow_flagging="never" ).launch()