最小二乘支持向量机在混沌时间序列中的应用_尹华.pdfVIP

28 2 2011 2 : 1006 - 9348( 2011) 02 - 0225- 03 尹 华, 吴 虹 (, 34 1000) : , , ( GA_LSSVM ) GA _LSSVM , , , Lorenz , GA _LSSVM GA _LSSVM , , : ; ; ; : TP309 : B Study onTime SeriesForecastingBased on LeastSquares SupportVectorMachine Y N Hua, WU H ong ( Co llege ofM a them atics and Computer Sc ience, A nnan N orm a lU n iversity, G uangzhou J iangx i 341000, Ch ina) ABSTRACT: Op tmi ization of the em bedd ing d mi ension, de lay tmi e and suppo rt vec to r machines param eters is a key prob lem in the pred ic tion of chao tic tmi e series based on phase space reconstruction and support vector m achines. n this paper, an a lgor ithm for optmi izing suppo rt vector m achine tmi e series forecasting m ode l param ete rs is proposed. F irstly, tmi e ser ies param eter is dec ided by C - C and G - P a lgo rithm. Second ly, the param eters of LSSVM is op ti- m ized by genetic a lgo rithm. T he result show s tha t the proposed algor ithm is accurate and the computational com plex i- ty is low. t is a good param eters optmi izing a lgo rithm. KEY ORDS: Support vecto r m ach ine( SVM ); T mi e ser ies; G ene tic a lgo rithm; O ptmi izing , 1 , , [ 1] , , , , 2 [ 2] [ 3] [ 4] , , , 2. 1 , , , ; , , , ( LSSVM )