Analysis of Water Quality Index for Indian Rivers Using AI Techniques

View at publisher Summarise with VIRAT
Authors: Neha Sharma, Abhineet Anand
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Assessment of water quality is essential to monitor the environment and public health. This chapter is intended for Indian river water quality classification through artificial intelligence. In combination, preprocess and augment an extensive range of microbiological, environmental, and physicochemical features with engineered features such as log-transformed coliform indicators, the deviation in pH, and biochemical oxygen demand/dissolved oxygen. The deep learning architectures (Multilayer Perceptron, TabNet, FT-Transformer), conventional machine learning techniques (Logistic Regression, Random Forest, and Gradient Boosting), and a hybrid ensemble (MLP + GB) are used. The classifier evaluation methodology provided the comparison by accuracy, precision, recall, F1-score, and log-loss. The hybrid ensemble outperformed any of the single models with the best log-loss (0.15), accuracy 0.97, precision 0.96, and F1-score 0.94. These results exceed those of previous studies and redefine water quality categories. Integration of complementary artificial intelligence (AI) paradigms for precise and intelligible Water Quality Index (WQI) prediction enhances the decision-support system for sustainable water resource management in India and elsewhere.

Fields of Study & Subjects

Water Quality and Pollution Assessment Hydrological Forecasting Using AI Water Quality Monitoring Technologies

Publication details

Journal / source
CRC Press eBooks
Pages
45–62
Publication date
31 March 2026
Language
EN
Document type
Book chapter