股票价格累积概率分布

来源：互联网发布：手机腾讯软件中心编辑：程序博客网时间：2024/04/29 20:14

sample.txt为上证指数，读入sas.

filename in 'c:\sample.txt';data a;   infile in  dlm='09'x;  input id   date :  yymmdd10.  (x1-x7) (: $12.) ;  format date yymmdd10.;proc print data=a;run;

进一步研究其开盘价趋势图：

data spprice( keep=price lprice );   set a;   price=input(x2,8.);   lprice=log(price);run;ods graphics on;proc timeseries data=spprice out=newprice plot=series;  var price lprice;run;ods graphics off;

当然还可以看看其描述性统计和简单检验的结果：

proc univariate data=spprice normal plot;   var price lprice;run;

试图查找价格指数的分布情况

1.各个价格出现比率

proc freq data=spprice ;   table price / noprint out=spfreq(keep=price percent);run;

2.将比率看做价格出现的概率，试图求出价格的累积分布密度

proc freq data=spprice ;   table price / noprint out=spfreq(keep=price percent);run;data  _null_;   if 0 then set spfreq nobs=n;   if _n_=1 then call symput('n',compress(put(n,8.0)));   stop;run;;data rndprice (keep=price) spcdf(keep=price spprob);   set spfreq end=endofprc;   array prices(&n) prc1-prc&n;   array prob(&n) prb1-prb&n;   retain prc1-prc&n prb1-prb&n;   prices(_n_)=price;   prob(_n_)=percent/100;if endofprc then do;   spprob=0;   do i=1 to dim(prices);      price=prices(i);   spprob=spprob+prob(i);   output ;   end;   do j=1 to 1000;      price=prices(rantbl(123,of prob(*)));      output ;   end; end; run;proc freq data=rndprice;   table price / noprint out=rndcdf(keep=price percent);run;data rndcdf2(keep=price rndprob);   set rndcdf;   retain rndprob 0;   rndprob=rndprob+percent/100;run;

作出累积分布密度图

data both;   set spcdf(rename=(spprob=prob))       rndcdf2(rename=(rndprob=prob) in=rnd);   retain sample 'theoretical';   if rnd then sample='random';   output;run;proc gplot data=both;  plot prob*price=sample;run;quit;

http://blog.renren.com/blog/220869953/729348859?from=fanyeNew