Prediction model of water consumption using least square support vector machines optimized by hybrid intelligent algorithm

Accurate forecasting of water consumption has been one of the most important issues in water supply system. Because of the high generalization performance and the ability to model non-linear relationships, least square support vector machines(LS-SVM) has been successfully employed to solve water consumption forecasting problems over the past few years. However, the practical use of LS-SVM is limited due to its set of parameters to be defined by the user. For this reason, this paper presents a LS-SVM parameter optimization approach based on genetic algorithms and particle swarm optimization(hybrid intelligent algorithm). It makes use of PSO algorithm characteristics such as parallel property and the global convergence performance to avoid the local optimum, and uses the evolution idea of genetic algorithm such as crossover and mutation operations to improve the speed of searching for the global optimization. At the same time, a deterministic searching algorithm is embedded to improve its optimization performance. The application in water consumption forecasting showed, this LS-SVM optimized by hybrid intelligent algorithm achieved better forecasting result.

[1]  Johan A. K. Suykens,et al.  Automatic relevance determination for Least Squares Support Vector Machines classifiers , 2001, ESANN.

[2]  J. Weidong.  Selection of Kernel Functions and Parameters for Support Vector Machines in System Identification , 2006 .

[3]  Yunqian Ma,et al.  Practical selection of SVM parameters and noise estimation for SVM regression , 2004, Neural Networks.

[4]  Johan A. K. Suykens,et al.  Least Squares Support Vector Machine Classifiers , 1999, Neural Processing Letters.

[5]  Yong Mao,et al.  Parameters selection in gene selection using Gaussian kernel support vector machines by genetic algorithm. , 2005, Journal of Zhejiang University. Science. B.

[6]  S. Sathiya Keerthi,et al.  Efficient tuning of SVM hyperparameters using radius/margin bound and iterative algorithms , 2002, IEEE Trans. Neural Networks.

[7]  Cheng Chuntian.  Application of support vector machine method to long-term runoff forecast , 2006 .

[8]  Sayan Mukherjee,et al.  Choosing Multiple Parameters for Support Vector Machines , 2002, Machine Learning.

[9]  Ryohei Nakano,et al.  Optimizing Support Vector regression hyperparameters based on cross-validation , 2003, Proceedings of the International Joint Conference on Neural Networks, 2003..

[10]  S. Sathiya Keerthi,et al.  Evaluation of simple performance measures for tuning SVM hyperparameters , 2003, Neurocomputing.

[11]  Zhou Xiao-bo,et al.  Parameters selection in gene selection using Gaussian kernel support vector machines by genetic algorithm , 2008, Journal of Zhejiang University Science B.