Fayrouz Abd Alkareem Hadi, Lariyah Mohd Sidek, Gasim Hayder Ahmed Salih, H. Basri, S. S. Sammen, Norlida Mohd Dom, Zaharifudin Muhamad Ali, Ali Najah Ahmed
{"title":"Machine learning techniques for flood forecasting","authors":"Fayrouz Abd Alkareem Hadi, Lariyah Mohd Sidek, Gasim Hayder Ahmed Salih, H. Basri, S. S. Sammen, Norlida Mohd Dom, Zaharifudin Muhamad Ali, Ali Najah Ahmed","doi":"10.2166/hydro.2024.208","DOIUrl":null,"url":null,"abstract":"\n \n The overarching goal of this research article is to examine the significant roles and substantial practicalities of artificial intelligence models in leading high-performance and accurate flood forecasting procedures. The Dungun River served as a case study. The forecasting procedure was implemented for two time spans: (1) the present period (1986–2000) and (2) the near future interval (2020–2030). Advanced machine learning algorithms engaged in this process were logistic regression, K-nearest neighbors, support vector classifier, Naive Bayes, decision tree, random forest, and artificial neural network. The results revealed that between 1986 and 2000, there would be an average of 18–55 floods around the Dungun River. Floods occurred rarely before 1985. Floods have been common since 2000. There have been about 35 floods annually on average since 2000. Meanwhile, it is predicted that between 2020 and 2030, the number of flooding events will grow due to climate change impacts on the Dungun River Basin. The maximum frequency of flooding was measured at 110 occurrences at a rainfall of 250 mm. The accuracy of the random forest was 75.61%, followed by the K-nearest neighbor at 73.17%. The accuracy of the logistic regression was the lowest (48.78%). Overall, the artificial neural networks models had a satisfying mean accuracy of 90.85%.","PeriodicalId":507813,"journal":{"name":"Journal of Hydroinformatics","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2024-02-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Hydroinformatics","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.2166/hydro.2024.208","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
The overarching goal of this research article is to examine the significant roles and substantial practicalities of artificial intelligence models in leading high-performance and accurate flood forecasting procedures. The Dungun River served as a case study. The forecasting procedure was implemented for two time spans: (1) the present period (1986–2000) and (2) the near future interval (2020–2030). Advanced machine learning algorithms engaged in this process were logistic regression, K-nearest neighbors, support vector classifier, Naive Bayes, decision tree, random forest, and artificial neural network. The results revealed that between 1986 and 2000, there would be an average of 18–55 floods around the Dungun River. Floods occurred rarely before 1985. Floods have been common since 2000. There have been about 35 floods annually on average since 2000. Meanwhile, it is predicted that between 2020 and 2030, the number of flooding events will grow due to climate change impacts on the Dungun River Basin. The maximum frequency of flooding was measured at 110 occurrences at a rainfall of 250 mm. The accuracy of the random forest was 75.61%, followed by the K-nearest neighbor at 73.17%. The accuracy of the logistic regression was the lowest (48.78%). Overall, the artificial neural networks models had a satisfying mean accuracy of 90.85%.