Application of Machine Learning for the Prediction of Drinking Water Quality at the Micaela Bastidas National University of Apurímac

Abstract

This  research  project  was  transcribed  to  present  to  the FERCYT Science and Technology Fair contest of 2024, in the scientific poster category, organized by the Micaela Bastidas University of Apurímac, the research project aims to evaluate the quality of drinking water that arrives at the Micaela Bastidas de Apurímac National University (UNAMBA) for this purpose, predictive learning models such as Neural Networks (RN) and Systems will be used. Adaptive Neuro-Diffuse (ANFIS). Data will be collected at the university, analyzed with water quality indicators. The models are corroborated by using metrics such as RMSE, which is the mean square error, MAE, which is the mean absolute error, and R², which is the coefficient of determination. The statistical tests that will be used are T student to evaluate significant differences in the quality parameters before and after fitting predictive models and ANOVA to compare the performances of RN and ANFIS on different water quality parameters. It is expected that the results will allow the implementation of real-time monitoring systems, contributing to the improvement of water quality in the institution.