Water demand prediction using machine learning: a review / Norashikin Nasaruddin ... [et al.]

• Main Authors: Nasaruddin, Norashikin (Author), Zakaria, Shahida Farhan (Author), Ahmad, Afida (Author), Ul-Saufie, Ahmad Zia (Author), Mohamaed Noor, Norazian (Author)
• Format: Book
• Published: 2021.
• Subjects: Conference or Workshop Item; PeerReviewed
• Online Access: Link Metadata
• Notes: https://ir.uitm.edu.my/id/eprint/56176/1/56176.pdf
• Keywords: T Technology (General); Technological innovations

Water is important and critical sources of life. Even though some countries enjoy tropical weather year-round with plenty of water resources like Malaysia, they are still facing scarcity issue. Water demand is influenced by various factors such as population, climate change and water utilization. This study reviews 45 Scopus articles from year 2015 to 2021 on predicting water demand using Machine Learning (ML) methods which include: neural network, random forest, decision tree, and hybrid optimisation models. The summary of ML methods on the evaluation of their performance in water demand prediction is identified by a comprehensive analysis of the literature. The narrative search of the most relevant literature is classified according to method, prediction type, prediction variables and accuracy rate. The review identified several machine learning methods that are commonly used which include decision tree, neural network, random forest and hybrid method. In conclusion, the study reports that the accuracy of the method varies according to types of prediction variables used.