Model Comparison for Water Quality Predictions: Manual vs IoT-Based Approaches

This project develops a water quality prediction system using machine learning and IoT devices toassess and classify water quality accurately, efficiently, and effectively. IoT sensors are deployed tocollect real-time data on various key water quality parameters, including pH, temperature, andconductivity, etc. Data is futher comprehensively analyzed. Missing values are preprocessed, and waterquality indices such as the National Sanitation Foundation Water Quality Index (NSFWQI) are applied.Various machine learning models, including linear regression, decision tree regressor, support vectorregressor, random forest regressor and gradient boosting regressor, are utilized to predict water quality.These models are evaluated using performance metrics such as Mean Absolute Error (MAE), MeanSquared Error (MSE), Root Mean Squared Error (RMSE), and R² score. The project highlights how AIand IoT can enhance water quality assessment for household, industrial, and environmental applicationsby enabling real-time monitoring and predictive analysis.