Study and Application of Slope Displacement Time Series Forecast Based on CO-WLSSVM

Abstract:

A multi-dimension admissible support vector wavelet kernel function is introduced and the model of wavelet least square support vector machine (WLSSVM) is optimized by chaos optimization (CO), which is named as wavelet least squares support vector machine based on chaos optimization (CO-WLSSVM).The optimized model improves the forecasting precision depending multi-dimension interpolation character and sparse change character of the wavelet and quick convergence to the optimum solution of the chaos optimization. The CO-WLSSVM is applied to forecast the displacement of left side bank of slope in first-stage hydroelectric station of Jinping. The result shows that the model possesses higher precision of forecasting.

• Periodical: Advanced Materials Research (Volumes 594-597)
• Pages: 2932-2935
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.594-597.2932
• Online since: November 2012
• Authors: Fei Xu, Ke Wang
• Keywords: Chaos Optimization, Least Squares Support Vector Machines (Ls-SVM), Slope Displacement, Time Series Forecast, Wavelet Kernel Function

References

[1] X.T. Feng: Introduction of Intelligent Rock Mechanics, edit by Science Press, pp.111-116. (In Chinese)

[2] J.A.K. Suykens, T.G. Van and M.B.De: Least squares support vector machine, edit by World Scientific(2002).

[3] L. DAUBECHIES: IEEE Trans on In formation Theory,Vol.36(1990),p: 961-1 005.

[4] Q.Zhang: IEEE Trans on Neural Networks,Vol.3(1992),pp.889-898.

[5] Z.Y. Song, J.J.Li: Engineering Journal of Wuhan University, Vol.40(2007),pp.53-57. (In Chinese)