经济波动谱分析方法的比较研究-控制与决策.pdfVIP

第18 卷第3 期 　控　制　与　决　策 2003 年5 月 Vol. 18 No . 3 　Control and Decision　 　May 2003 : 1001-0920( 2003) 03-0351-04 徐　梅, 梅世强 ( , 300072) : 为解决我国经济序列长度较短及各主周期分量难以同时准确分 的问题, 提出加窗周期图谱 估计与最大熵谱估计相结合的谱估计方法, 并将各主要周期分量单独分 , 且通过对序列进行三角函数 拟合来确定周期长度准确值。从谱估计曲线、主周期分量的分 和周期长度的准确性3 个方面与 一般的 谱分析方法进行了比较, 说明了该方法的可行性和有效性。 : 经济波动; 加窗周期图谱估计; 最大熵谱估计; 分 ; 比较 : F224: A Comparative study of spectral analysis method of business fluctuation X U M ei, MEI S hi-qiang ( , , 300072, ) School of Manag m nt T ianjin U niv rsity T ianjin China A bstract: A sp ctral stimation of combining p riodogram w indow ing sp ctral stimation and max imum ntropy sp ctral stimation is sugg st d to solv th probl m that th l ngth of conom ic tim s ri s in China is short. A m t hod of d t rmining th accurat valu of p riod l ngt h of ach m ain hiding p riod is put forw ard t o d al w it h th probl m that ach main hiding p riod is difficult to b r solv d accurat ly at th sam t im , Th m thod r solv s hiding p riod on by on and fitting data w ith triangular function. Th m thod is show n to b ff ct iv and r asonabl by comparing w it h ordinary sp ctral analysis m thod in thr asp cts of sp ctral stimat ion curv , r solving of main hiding p riods and accuracy of p riod l ngth to show that it is ff ctiv and r asonabl . Key words : Busin ss fluctuation; P riodogram window ing sp ctral stimat ion; Max imum ntropy sp ctral stimation; R solv ; Comparison 1引　　 言 {xt} , t {X }