Problem Statement: Heart defects are among the leading causes of mortality worldwide. Early detection using machine learning can assist healthcare professionals in identifying high-risk patients and improving diagnostic accuracy.

Objective: To build a predictive model that classifies whether a patient is likely to have a heart defect based on clinical and physiological parameters.

Dataset: Patient medical records Features include: Age Sex Blood pressure(Systolic & Diastolic) Cholesterol level Glucose level Fasting blood sugar Maximum heart rate Prevalent Stroke & Hypertension

Methodology: Data Preprocessing Handling missing values using Iterative Imputer Feature scaling using Robust Scaler Encoding categorical variables Exploratory Data Analysis (EDA) Feature importance visualization

Model Training: Random Forest

Hyperparameter Tuning: GridSearchCV RandomSearchCV Bayesian Optimization

Model Evaluation: Accuracy Precision, Recall, F1-score Confusion Matrix

Results: Random Forest achieved 84.98% accuracy Bayesian Optimization improved model accuracy and generalization

Tech Stack: Python NumPy, Pandas Scikit-learn Matplotlib, Seaborn

Use Cases: Clinical decision support systems Preventive healthcare analytics Medical research assistance