Predictive Water Quality Modeling Using ARIMA and VAR for Locations of Krishna River, Andhra Pradesh, India

Abstract

AbstractThis paper investigates the recital use of machine learning techniques that include vector autoregression (VAR) and autoregression integrated mean average (ARIMA) for the Krishna River of Andhra Pradesh, India. For modeling, water parameters are collected from the Central Pollution Control Board, India, and Andhra Pradesh Pollution Control Board. The process applied in this paper had shown tensing during the code generation and transfer to the Jupyter notebook functioning. Comparing both models, VAR outperformed ARIMA for predicting the next six progressive values of the locations. After examining the models’ results, the ARIMA got less than 0.5, and RSME of VAR had 0.95; the VAR is notified to be an accurate model for predicting the next six consecutive values. In parallel to the machine learning process, the respective study location samples are collected and tested in the laboratory to compare the results derived and predicted.KeywordsARIMAVARJupyter notebookKrishna RiverArtificial intelligent