An enhancement of sliding window algorithm for rainfall forecasting

• Main Authors: Azahar, Siti Nor Fathihah, Othman, Mahmod, Saian, Rizauddin
• Format: Conference or Workshop Item
• Language: English
• Published: 2017
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://repo.uum.edu.my/22780/1/ICOCI%202017%2023-28.pdf; http://repo.uum.edu.my/22780/; http://icoci.cms.net.my/PROCEEDINGS/2017/Pdf_Version_Chap01e/PID29-23-28e.pdf

Various rainfall forecasting models or techniques are presented by researchers to obtain the best result of forecasting. Despite of various techniques and methods, not all of them produce the satisfy result of rainfall forecasting. Therefore, this study proposed a forecasting rainfall method based on sliding window algorithm (SWA), in order to obtain the best rainfall prediction value. The problem statements in this study are related to unsatisfactory accuracy rainfall output on previous study of SWA. Hence, SWA is enhanced in order to produce highly accurate prediction rainfall. The proposed method is tested by using three different rainfall gauge station data that are taken from the Department of Drainage Irrigation (DID), Perlis, Malaysia. Then, the rainfall forecasting result is validated by using mean square error (MSE), and relative geometric root mean squared error (relative GRMSE). The validation analysis shows that the proposed method has a higher forecasting accuracy than the previous method of sliding window algorithm.