ziqubusibusiso-svg/Predicting-the-functionality-of-water-points-in-Tanzania-using-machine-learning

Machine learning project # Predicting-the-functionality-of-water-points-in-Tanzania-using-machine-learning This project compares the performance of Random Forest and XGBoost models in predicting water point functionality, with a particular focus on how model selection impacts predictive accuracy. To address class imbalance in the dataset, SMOTE (Synthetic Minority Over-sampling Technique) is applied, significantly improving model performance. Experimental results show that Random Forest combined with SMOTE outperforms XGBoost, achieving an accuracy of approximately 70%. Model evaluation using confusion matrices demonstrates Random Forest’s effectiveness in correctly identifying functional and non-functional water points. Feature importance analysis further provides insights into the key factors influencing water point functionality. This repository contributes to understanding the suitability of different machine learning models for infrastructure reliability prediction in developing contexts and highlights the potential of predictive analytics to support sustainable water access initiatives.