UMP Institutional Repository

Integration of GWO-LSSVM for time series predictive analysis

Zuriani, Mustaffa and Mohd Herwan, Sulaiman and Bariah, Yusob and Ernawan, Ferda (2016) Integration of GWO-LSSVM for time series predictive analysis. In: 4th IET Clean Energy and Technology Conference, CEAT 2016, 14-15 November 2016 , Kuala Lumpur, Malaysia. pp. 1-5., 2016 (CP688). ISBN 978-1-78561-238-1

[img] Pdf
Integration Of GWO-LSSVM For Time Series Predictive Analysis.pdf - Published Version
Restricted to Repository staff only

Download (512kB) | Request a copy


The emergence of Statistical Learning Theory (SLT) based algorithm namely Least Squares Support Vector Machines (LSSVM) has evidenced its efficacy in solving regression and classification problems. In this study, LSSVM is employed as a predictor to predict water level. Accurate water levels modeling and prediction is vital for safety especially during the monsoon season. It is worth noting that, mitigating the effects of floods can be accomplished by using either structural or non-structural measures, or by a combination of both. Structural measures incorporated of engineering works, such as channelization or flood reservoirs which changes the shape of the flood hydrograph. On the other hand, non-structural measures such as flood warning scheme is intended to reduce the economic losses in a flood situation. In most cases, disaster prevention operation such flood warning scheme proved to be more efficient in mitigating the effects of major floods than structural measures. Thus, for this study, a hybrid algorithm of LSSVM with one of the recent bio-inspired optimization algorithm, namely Grey Wolf Optimizer (GWO-LSSVM) is presented for water level prediction. The GWO is utilized to optimize the parameters of LSSVM. The feasibility of GWO-LSSVM is compared against LSSVM optimized by Firefly Algorithm (FA-LSSVM) and single GWO. Findings of the study demonstrated that the GWO-LSSVM able to provide competitive results and able to perform well for the problem under study.

Item Type: Conference or Workshop Item (Speech)
Additional Information: Indexed by Scopus
Uncontrolled Keywords: Grey wolf optimizer; Least squares support vector machines; Optimization; Prediction
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Faculty/Division: Faculty of Computer System And Software Engineering
Faculty of Electrical & Electronic Engineering
Depositing User: Dr. Zuriani Mustaffa
Date Deposited: 12 Sep 2018 07:29
Last Modified: 12 Sep 2018 07:29
Download Statistic: View Download Statistics

Actions (login required)

View Item View Item