❤️ Heart Attack Prediction Using Logistic Regression and Gaussian Naive Bayes

📌 Project Overview

This project predicts the possibility of a heart attack using classification models Logistic Regression and Naive Bayes.
It uses patient health data from the Cleveland Heart Disease dataset to classify whether a person has a lower or higher risk of heart attack.

The project focuses on:

• Data loading and preprocessing
• Training multiple classification models
• Evaluating performance using Accuracy, Precision and Recall

---

🎯 Objective

To build and compare machine learning models that can:

• Analyze medical attributes of patients
• Predict heart attack possibility (0 or 1)
• Provide a simple and interpretable healthcare prediction system

---

📊 Dataset Information

• Dataset: Cleveland Heart Disease Dataset
• File Used: heart.csv
• Target Column: target
  • 0 → Less chance of heart attack
  • 1 → More chance of heart attack

---

🧬 Features Used

The model is trained using the following medical attributes:

Feature	Description
age	Age of the patient
sex	Gender (1 = male, 0 = female)
cp	Chest pain type
trestbps	Resting blood pressure
chol	Serum cholesterol (mg/dl)
fbs	Fasting blood sugar > 120 mg/dl
restecg	Resting ECG results
thalach	Maximum heart rate achieved
exang	Exercise induced angina
oldpeak	ST depression induced by exercise
slope	Slope of peak exercise ST segment
ca	Number of major vessels
thal	Thalassemia type

---

🛠️ Technologies Used

• Python
• Pandas – data handling
• Matplotlib & Seaborn – data visualization
• Scikit-learn – machine learning

---

🧠 Machine Learning Models Used

Logistic Regression

• Used for binary classification
• Outputs class predictions (0 or 1)
• Simple, efficient, and interpretable for healthcare data

model = LogisticRegression(max_iter=1300)
model.fit(X_train, y_train)

Gaussian Naive Bayes

• Based on Bayes Theorem
• Assumes features follow a Gaussian distribution
• Fast and effective for probabilistic classification

gnb_model = GaussianNB()
gnb_model.fit(X_train, y_train)

---

🔄 Workflow

• Load dataset using Pandas
• Separate features (X) and target (y)
• Split data into training and testing sets (80% / 20%)
• Train Logistic Regression and Gaussian Naive Bayes models
• Make predictions on test data
• Evaluate and compare model performance

---

📈 Model Evaluation

The model is evaluated using Accuracy, Precision and Recall metrics.

print("accuracy: ", accuracy_score(y_test, y_pred) * 100, "%")
print("precision: ", precision_score(y_test, y_pred) * 100, "%")

print("recall score: ", recall_score(y_test, y_pred))
print("precision: ", precision_score(y_test, y_pred))
print("accuracy_score: ", accuracy_score(y_test, y_pred))

🔹 Metrics Explained

Accuracy: Measures overall correctness of predictions

Precision: Measures how many predicted positive cases are actually positive (important in medical diagnosis)

Recall: Proportion of actual positive cases correctly identified 📌 Recall is especially important in healthcare to minimize false negatives.

---

✅ Results and Comparison

Both Logistic Regression and Gaussian Naive Bayes models were evaluated on the test dataset using standard classification metrics.

🔹 Logistic Regression Performance

Accuracy: 0.8525 Precision: 0.8710 Recall: 0.8438

🔹 Gaussian Naive Bayes Performance

Accuracy: 0.8689 Precision: 0.9000 Recall: 0.8438

📈 Model Comparison Summary

Metric	Logistic Regression	Gaussian Naive Bayes
Accuracy	0.8525	0.8689
Precision	0.8710	0.9000
Recall	0.8438	0.8438

Interpretation of Results

• Gaussian Naive Bayes achieves higher accuracy and precision, indicating fewer false positive predictions.

• Both models show identical recall, meaning they are equally effective at identifying actual heart attack cases.

• Logistic Regression offers better interpretability, while Gaussian Naive Bayes provides slightly better predictive performance.

📌 Overall, Gaussian Naive Bayes performs marginally better on this dataset, while Logistic Regression remains a strong and interpretable baseline.

---