Scholarlink

Groundwater quality prediction is vital for managing scarce water resources in regions like Vietnam and India, where contamination poses environmental and health risks. This paper introduces AI-LGBM, a novel model that integrates Mutual Information-based Feature Selection (MIFS) and employs Particle Swarm Optimization (AIO) for hyperparameter tuning within LightGBM to predict groundwater quality. Evaluated on datasets from Vietnam and India, AI-LGBM achieved superior performance, attaining an accuracy of 98.0%, precision of 98.10%, recall of 98.00%, and F1-score of 98.00%, outperforming models like Random Forest (95.00% accuracy), XGBoost (94.00%), and SVM (93.00%). Leveraging explainable AI techniques such as LIME and SHAP, AI-LGBM provides accurate predictions and valuable insights into key features influencing water quality, enhancing its utility for decision-making in water resource management.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Using Auto Immune LightGBM (AI-LGBM) for Prediction of Ground Water Quality in Vietnam and Indian Regions

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Michael Omar,
  • Nguyen Long Giang,
  • Tran Thi Ngan,
  • Nguyen Hong Tan,
  • Nguyen Thu Van

摘要

Groundwater quality prediction is vital for managing scarce water resources in regions like Vietnam and India, where contamination poses environmental and health risks. This paper introduces AI-LGBM, a novel model that integrates Mutual Information-based Feature Selection (MIFS) and employs Particle Swarm Optimization (AIO) for hyperparameter tuning within LightGBM to predict groundwater quality. Evaluated on datasets from Vietnam and India, AI-LGBM achieved superior performance, attaining an accuracy of 98.0%, precision of 98.10%, recall of 98.00%, and F1-score of 98.00%, outperforming models like Random Forest (95.00% accuracy), XGBoost (94.00%), and SVM (93.00%). Leveraging explainable AI techniques such as LIME and SHAP, AI-LGBM provides accurate predictions and valuable insights into key features influencing water quality, enhancing its utility for decision-making in water resource management.