Day 9 & Day 10 – Heart Disease Prediction | NTI Zagazig
تفاصيل العمل
1. Data Cleaning Dataset: heartdiseaseuci.csv (920×16) → cleaned to 675 rows Removed unnecessary columns (id) Handled missing values using domain-aware strategies Converted categorical and boolean features to numeric 2. Feature Engineering Created domain-driven features: Age Risk Groups Blood Pressure Risk Cholesterol Risk Heart Rate Reserve ST Depression Indicator Used ColumnTransformer + Pipeline → 32 engineered features 3. Modeling & Evaluation Initial model: KNN + GridSearchCV → Accuracy ≈ 60% Challenge: imbalanced dataset & rare classes Next step: Random Forest / Ensemble Models → expected Accuracy 87–90%
مهارات العمل
بطاقة العمل
طلب عمل مماثل
