Hydroclimatic Data Prediction using a New Ensemble Group Method of Data Handling Coupled with Artificial Bee Colony Algorithm; [Ramalan Data Hidroklimatik menggunakan Kaedah Pengendalian Data Kumpulan Ensembel Baharu Digandingkan dengan Algoritma Koloni Lebah Buatan]

• Published in: Sains Malaysiana
• Main Author: Badyalina B.; Mokhtar N.A.; Jan N.A.M.; Marsani M.F.; Ramli M.F.; Majid M.; Ya’Acob F.F.
• Format: Article
• Language: English
• Published: Penerbit Universiti Kebangsaan Malaysia 2022
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85140344916&doi=10.17576%2fjsm-2022-5108-24&partnerID=40&md5=f392c793ba8d2efdf5e1f926e8659b30

Linear regression is widely used in flood quantile study that consists of meteorological and physiographical variables. However, linear regression does not capture the complex nonlinear relationship between predictor and target variables. It is rare to find a hydrological application using the group method of data handling (GMDH) model, artificial bee colony (ABC) algorithm, and ensemble technique, precisely predicting ungauged sites. GMDH model is known to be an effective model in complying with a nonlinear relationship. Therefore, in this paper, we enhance the GMDH model by implementing the ABC algorithm to optimize the parameter of partial description GMDH model with some transfer functions, namely polynomial, radial basis, sigmoid and hyperbolic tangent function. Then, ensemble averaging combines the output from those various transfer functions and becomes the new ensemble GMDH model coupled with the ABC algorithm (EGMDH-ABC) model. The results show that this method significantly improves the prediction performance of the GMDH model. The EGMDH-ABC model satisfies the nonlinearity in data to produce a better estimation. Also, it provides more robust, accurate, and efficient results. © 2022 Penerbit Universiti Kebangsaan Malaysia. All rights reserved.