A Systematic Review On The Trends, Progresses, And Challenges In The Application Of Artificial Intelligence In Water Quality Assessment And Monitoring In Nigeria

Abstract

In recent decades, machine learning (ML) artificial intelligence has found wide application in water quality monitoring and prediction due to the increasing complexity of water quality data. This complexity has been attributed to the global upsurge in anthropogenic activities and climatic variations. It is therefore critical to identify the most accurate and suitable ML model for water quality prediction. In this study, a systematic literature review (SLR) was carried out to explore the trend and progress in the application of ML models in water quality monitoring and prediction in Nigeria from 2003-2024. A comprehensive review of the effectiveness of advanced ML models as well as the gaps in their application in the area of water quality assessment and monitoring was also carried out using the PRISMA-P meta-analysis technique. Forty publications were used to perform bibliographic analysis and visualization using the VOS viewer software. The study found that globally, the use of hybrid ML models in water quality prediction has not been well explored; a majority of the prediction has been based on the use of artificial neural networks (ANN). Among the ANN algorithms, the adaptive neuro-fuzzy inference system (ANFIS), and Wavelet-Adaptive Neural Fuzzy Interference System (W-ANFIS) hybrid models are the most accurate in prediction; with temperature, dissolved oxygen (DO), pH, electrical conductivity (EC), and total dissolved solids (TDS) among the most frequently predicted parameters. Nigeria is grossly lagging in the application of ML in water quality prediction. This limitation is largely attributed to inadequate data on environmental monitoring. It is critical therefore for future water quality monitoring and prediction studies in Nigeria to take advantage of the rapidly evolving field of machine learning; with more emphasis placed on the hybridized machine learning algorithms