论文信息 - Research on the Optimized Support Vector Regression Machines Based on the Differential Evolution Algorithm

Research on the Optimized Support Vector Regression Machines Based on the Differential Evolution Algorithm

The Support Vector Regression machine (SVR) is an effective tool to solve the problem of nonlinear prediction, but its prediction accuracy and generalization performances depend on the selection of parameters greatly. And the parameters selection is a procedure of global optimization search. Since the Differential Evolution (DE) population-based algorithm is a real- coding optimal algorithm with powerful global searching capacity, a hybrid model of DE-SVR based on the standard SVR model and DE algorithm is proposed in this paper. And then, the new hybrid implementation was applied to the short range regression prediction of the chaotic time series. At last, the experiment results showed the effectiveness of this approach and the better performance in searching time, compared with the conventional parameters searching approach of grid algorithm. algorithm has been achieved fine application effects in neural network training, filter design, cluster analysis. And this paper proposed a hybrid model based on the DE algorithm and the Support Vector Regression machine (SVR) model, which is called DE-SVR here. The proposed model is then applied to the short range regression forecast of the chaotic time series of Chens. The results of the proposed DE-SVR approach are compared with the results obtained by the conventional grid searching method and the PSO algorithm. Results presented show that the proposed DE-SVR approach appears to in higher prediction accuracy and faster searching.

[1] Fraser,et al. Independent coordinates for strange attractors from mutual information. , 1986, Physical review. A, General physics.

[2] R Hegger,et al. Improved false nearest neighbor method to detect determinism in time series data. , 1999, Physical review. E, Statistical physics, plasmas, fluids, and related interdisciplinary topics.

[3] Vladimir N. Vapnik,et al. The Nature of Statistical Learning Theory , 2000, Statistics for Engineering and Information Science.

[4] Cheng-Lung Huang,et al. A GA-based feature selection and parameters optimizationfor support vector machines , 2006, Expert Syst. Appl..

[5] Nello Cristianini,et al. An Introduction to Support Vector Machines and Other Kernel-based Learning Methods , 2000 .

[6] Rainer Storn,et al. Differential Evolution – A Simple and Efficient Heuristic for global Optimization over Continuous Spaces , 1997, J. Glob. Optim..

[7] Wu,et al. Parameter Selection Method for SVM with PSO , 2006 .

[8] Sadasivan Puthusserypady,et al. Chaotic time series prediction and additive white Gaussian noise , 2007 .

[9] Yann LeCun,et al. Measuring the VC-Dimension of a Learning Machine , 1994, Neural Computation.

[10] Rainer Storn,et al. Differential Evolution-A simple evolution strategy for fast optimization , 1997 .

[11] Hahn-Ming Lee,et al. Model selection of SVMs using GA approach , 2004, 2004 IEEE International Joint Conference on Neural Networks (IEEE Cat. No.04CH37541).