IoT-based automated system for water-related disease prediction

Area of Science:

• Environmental Science
• Public Health
• Data Science

Background:

• Access to potable water is a fundamental right, yet 3.4 million people die annually from waterborne diseases, and 1.1 billion lack safe drinking water.
• Human activities contaminate water sources, making water responsible for an estimated 80% of illnesses.
• Industrialization and development have not solved the global water crisis, necessitating advanced solutions.

Purpose of the Study:

• To develop a real-time system for monitoring water quality and predicting waterborne diseases.
• To forecast long-term water quality trends using advanced machine learning techniques.
• To address the critical issue of water scarcity and contamination impacting global health.

Main Methods:

• Utilized a real-time West Bengal Pollution Control Board (WBPCB) dataset with 17 features.
• Proposed an Internet of Things (IoT)-based system for real-time data collection using Hybrid Adaptive Neuro-Fuzzy Inference System (H-ANFIS).
• Employed machine learning algorithms including Random Forest, XGBoost, AdaBoost for classification and Long Short-Term Memory (LSTM) for forecasting. Introduced TS-SMOTE for data augmentation.

Main Results:

• Classification models achieved high accuracy: Random Forest (99.66%), AdaBoost (99.64%), and XGBoost (99.52%).
• LSTM forecasting for the pH parameter yielded a Mean Squared Error (MSE) of 0.1631.
• The IoT system effectively gathered data in real-time and identified potential attacks.

Conclusions:

• The proposed IoT system and machine learning models demonstrate high efficacy in predicting waterborne diseases and forecasting water quality.
• The research offers a robust framework for ensuring a reliable supply of potable water and mitigating health risks.
• The novel TS-SMOTE approach enhances the performance of time-series analysis for water quality monitoring.