出于模型的须要,咱们的团队选择作一次因子分析,一般这部分在队伍中是会有同窗专门负责这块的,至于为何笔者就不在这里多说了。code
在MATLAB中封装了有关因子分析的方法--factoran
,读者能够经过help
命令来查看如何调用这个方法。get
须要读者注意的是,factoran方法并不会作数据规范化,因此读者须要本身来作这个操做。it
%数据单位化 TX_F_D=zscore(TX_D) %5,求和 SUM_F_D=TX_F_D+AZ_F_D+CA_F_D+NM_F_D %6,因子分析 [SUM_lambda,SUM_psi,SUM_T,SUM_stats,SUM_F]=factoran(SUM_F_D,6) %求取贡献率 SUM_Contribute=Factor_Contribute(SUM_lambda,index) %求各项因子的得分 ALL_F=Factor_F(TX_F_D,AZ_F_D,CA_F_D,NM_F_D,SUM_lambda,SUM_Contribute) %画图 subplot(2,2,1) plot(YEARS,ALL_F{1}(:,1),'r-',YEARS,ALL_F{2}(:,1),'g--',YEARS,ALL_F{3}(:,1),'b:',YEARS,ALL_F{4}(:,1)) xlabel('YEARS') ylabel('F1') legend('TX','AZ','CA','NM','Location','SouthEast') %画单个的图 figure plot(YEARS,ALL_F{1}(:,1),'r-',YEARS,ALL_F{2}(:,1),'g--',YEARS,ALL_F{3}(:,1),'b:',YEARS,ALL_F{4}(:,1)) xlabel('YEARS') ylabel('F1') legend('TX','AZ','CA','NM','Location','SouthEast') figure plot(YEARS,ALL_F{1}(:,2),'r-',YEARS,ALL_F{2}(:,2),'g--',YEARS,ALL_F{3}(:,2),'b:',YEARS,ALL_F{4}(:,3)) xlabel('YEARS') ylabel('F2') legend('TX','AZ','CA','NM','Location','SouthEast') figure plot(YEARS,ALL_F{1}(:,3),'r-',YEARS,ALL_F{2}(:,3),'g--',YEARS,ALL_F{3}(:,3),'b:',YEARS,ALL_F{4}(:,3)) xlabel('YEARS') ylabel('F3') legend('TX','AZ','CA','NM','Location','SouthEast') figure plot(YEARS,ALL_F{1}(:,4),'r-',YEARS,ALL_F{2}(:,4),'g--',YEARS,ALL_F{3}(:,4),'b:',YEARS,ALL_F{4}(:,4)) xlabel('YEARS') ylabel('F') legend('TX','AZ','CA','NM','Location','SouthEast')
因为结果有多种多样的,直接给出MATLAB的工做空间,有兴趣的读者能够自行下载。io
连接 密码:zffrast