Prediction of Water Quality Using SoftMax-ELM Optimized Using Adaptive Crow-Search Algorithm

Abstract

Water is a predominant source in the survival and development of all human lives. On top of all, predicting water quality is a significant one since water is essential in regulating our human body. In recent days, the advent of machine learning techniques has been supporting a lot in water quality prediction. Accordingly, Adaptive Crow Search Optimized SoftMax-Extreme Learning Machine (AdCSO-sELM) is proposed to improve the ELM performance by making the flight length adaptively with respect to the iterations. Here, the research novelty lies in making the CSOA parameters as a dynamic one which in turn provides promising ELM performance. Finally, the proposed AdCSO-sELM provides a superior accuracy of 96.54% for classifying water potability using the Kaggle dataset.