Pralhad P. Teggi <br/><br/>
Predicting Urban Vartur Lake Health: A Multi-context Machine Learning Approach to Water Quality Assessment 

 - p577–590 
<br/><br/> This study introduces a novel multi-context approach to water quality assessment by examining the factors influencing Chemical Oxygen Demand in Varthur Lake, Bangalore. Integrating traditional regression techniques with advanced machine learning models, this research analyzes eight water quality parameters collected over two years (July 2022–September 2024). Exploratory Data Analysis is employed to understand parameter distributions, relationships, and trends. Five distinct environmental contexts affecting COD are identified, and various machine learning models are evaluated for COD prediction within each context. Results indicate that Random Forest models yield superior accuracy, with Gradient Boosting offering a viable alternative. This multi-context framework provides a comprehensive understanding of Varthur Lake’s water health and offers a scalable approach for predictive water quality monitoring applicable to other urban water bodies. 
<br/><br/><label>Subjects--Facted Index Terms: </label><br/>Environmental Sciences <br/>Limnology <br/>Machine Learning <br/>Water and Health <br/>Water Quality and Water Pollution <br/>Water Treatment