时间序列数据_进行数据预测_如预测下一年度的销售额、预测股票价格的走势等_matlab

%% 三次指数平滑 clc,clear load('yx.mat') %原始数据以列向量的方式存放 yt=yx; n=length(yt); alpha=0.9; st1_0=mean(yt(1:3)); st2_0=st1_0; st3_0=st1_0; st1(1)=alpha*yt(1)+(1-alpha)*st1_0; st2(1)=alpha*st1(1)+(1-alpha)*st2_0; st3(1)=alpha*st2(1)+(1-alpha)*st3_0; for i=2:n st1(i)=alpha*yt(i)+(1-alpha)*st1(i-1); st2(i)=alpha*st1(i)+(1-alpha)*st2(i-1); st3(i)=alpha*st2(i)+(1-alpha)*st3(i-1); end st1=[st1_0,st1]; st2=[st2_0,st2]; st3=[st3_0,st3]; a=3*st1-3*st2+st3; b=0.5*alpha/(1-alpha)^2*((6-5*alpha)*st1-2*(5-4*alpha)*st2+(4-3*alpha)*st3); c=0.5*alpha^2/(1-alpha)^2*(st1-2*st2+st3); yhat=a+b+c; yhat=yhat';