Stunting Risk Classification

My Contribution

• Performed exploratory data analysis on demographic, growth, breastfeeding, and stunting status features.
• Inspected missing values, duplicate records, feature distributions, outliers, and correlations.
• Removed duplicate records and handled numerical outliers using the Interquartile Range (IQR) method.
• Encoded categorical features using LabelEncoder and standardized features with StandardScaler.
• Prepared the dataset for classification by separating the features and stunting status target.
• Trained and compared Logistic Regression, Random Forest, and XGBoost models.
• Evaluated model performance using confusion matrix, classification report, Accuracy, Precision, Recall, and F1-Score.
• Selected Logistic Regression as the main model based on its Accuracy and Recall for identifying positive cases.
• Performed hyperparameter tuning experiments for Logistic Regression and XGBoost using GridSearchCV.
• Developed an interactive Streamlit dashboard for dataset exploration, filtering, and model comparison.