Enhancing Water Quality Prediction Through Explainable Machine Learning

Akuthota, Pramod (2025) Enhancing Water Quality Prediction Through Explainable Machine Learning. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (804kB) | Preview
Preview	PDF (Configuration Manual) Download (403kB) | Preview

Abstract

The increasing importance of predicting water quality is a result of environmental degradation, urbanization, and the need for quick access to freshwater resources. Traditional laboratory methods for measuring water quality are inefficient, leading to a rise in demand for automated and data-driven means of assessing water quality. Machine learning (ML) and deep learning (DL) models are able to provide good predictive performance. However, their black-box nature makes it difficult to trust them in real-world applications. This paper seeks to fill that gap by integrating Random Forest with LIME in order to provide both high predictive accuracy and interpretability. The physicochemical characteristics of water are normalized using the SelectKBest based on ANOVA statistical analyses, and class imbalance is handled with SMOTE. The Random Forest model can learn how features interact with each other non-linearly. However, the interpretation of feature importance is not as clear to the end user as the output of LIME would indicate. Together, the integrated framework would enable reliable classification of water quality along with transparent decision support. This integrated framework demonstrated its ability to provide high-quality results across individual datasets from Canada, China, England, Ireland and the United States, including a score of R2 = 0.99 and a very high classification accuracy.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Moldovan, Arghir Nicolae UNSPECIFIED
Subjects:	G Geography. Anthropology. Recreation > GE Environmental Sciences Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	07 Sep 2026 08:32
Last Modified:	07 Sep 2026 08:32
URI:	https://norma.ncirl.ie/id/eprint/9842

Actions (login required)

View Item