Statistical analysis and prediction via neural networks of water quality in the Middle Paraíba do Sul (Rio de Janeiro State, Brazil) region in the period (2012–2022)

Abstract

The aim of this study is to accurately predict the water quality at these points over a decade through the combined use of statistical tools and artificial intelligence. This study brings the innovative use of neural networks implemented with the GRNN package of the R statistical software to predict the water quality of nine points on the Paraíba do Sul River with their appropriate metrics. After choosing the points to be studied, specific information about the river was taken from the INEA database and treated statistically using tools such as ANOVA, multiple regression, and artificial intelligence using the R software. After processing the historical data, the results were discussed, interpreted, and critically analyzed, which led to a conclusive analysis of the data. As a result, the predictive model for water quality using artificial neural networks was developed and showed high accuracy when validated with precise data, as indicated by the metrics presented. The results of this study not only improve understanding of the factors that influence water quality, but also offer practical guidelines for management and intervention policies, contributing to the preservation and recovery of water resources in the region.