<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>topic Re: how to repeatedly run regression for larger sets in SAS Programming</title>
    <link>https://communities.sas.com/t5/SAS-Programming/how-to-repeatedly-run-regression-for-larger-sets/m-p/78887#M17049</link>
    <description>&lt;HTML&gt;&lt;HEAD&gt;&lt;/HEAD&gt;&lt;BODY&gt;&lt;P&gt;I am no economist or financial expert but the following simulation should warn you of a potential problem when looking for linear relationships between autocorrelated time series :&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;/* Generate 100 simulated daily price series */&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;data test;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;format date yymmdd10. price 7.2;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;call streaminit(658487);&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;do stock = 1 to 100;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; price = 50 + 100 * rand('UNIFORM');&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; do date = '01jan2000'd to '31dec2004'd;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; price = price * (1 + rand("NORMAL", 0.0, 0.01));&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; output;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; end;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; end;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;run;&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;proc sort data=test; by date stock; run;&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;/* Make each simulated stock series into a table column */&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;proc transpose data=test out=table prefix=stock_;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;by date;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;id stock;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;var price;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;run;&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;/* Plot the first three series */&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;title "Three simulated price series";&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;ods graphics / antialiasmax=2000;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;proc sgplot data=table;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;series x=date y=stock_1 /lineattrs=(pattern=1);&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;series x=date y=stock_2 /lineattrs=(pattern=1);&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;series x=date y=stock_3 /lineattrs=(pattern=1);&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;xaxis type=time tickvalueformat=yymmdd10.;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;yaxis label="Price";&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;run;&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;/* Find the 5 best correlations for each series */&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;ods listing select none;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;proc corr data=table best=5 Pearson;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;var stock_:;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;ods output PearsonCorr=PC;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;run;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;ods listing select all;&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;options linesize=132;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;title "Best correlated price series";&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;proc print data=PC(drop=best1 R1 P1) noobs; run;&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;You will notice that far too many relations appear as highly significant among these totally unrelated time series. Looking at the plot should give you a feeling why this happens.&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt; &lt;IMG alt="SGPlot6.png" class="jive-image-thumbnail jive-image" src="https://communities.sas.com/legacyfs/online/3212_SGPlot6.png" width="450" /&gt;&lt;/P&gt;&lt;P&gt;PG&lt;/P&gt;&lt;/BODY&gt;&lt;/HTML&gt;</description>
    <pubDate>Sat, 16 Mar 2013 02:17:35 GMT</pubDate>
    <dc:creator>PGStats</dc:creator>
    <dc:date>2013-03-16T02:17:35Z</dc:date>
    <item>
      <title>how to repeatedly run regression for larger sets</title>
      <link>https://communities.sas.com/t5/SAS-Programming/how-to-repeatedly-run-regression-for-larger-sets/m-p/78886#M17048</link>
      <description>&lt;HTML&gt;&lt;HEAD&gt;&lt;/HEAD&gt;&lt;BODY&gt;&lt;P&gt;I have a file with 100 coloumns(each column is stock data for 5 years). I have to run OLS between every 2 stocks(proc reg with model statement), so output is going to be coefficients for 4950 (100*99/2) pair of stocks.&lt;/P&gt;&lt;P&gt;I can put proc reg in macro, but I have to call it 4950 times..or I can use matlab (with for loops) to make it less tedious.&lt;/P&gt;&lt;P&gt;Is there a better solution of doing this in SAS (besides macros). thank you&lt;/P&gt;&lt;/BODY&gt;&lt;/HTML&gt;</description>
      <pubDate>Fri, 15 Mar 2013 23:25:32 GMT</pubDate>
      <guid>https://communities.sas.com/t5/SAS-Programming/how-to-repeatedly-run-regression-for-larger-sets/m-p/78886#M17048</guid>
      <dc:creator>kashili</dc:creator>
      <dc:date>2013-03-15T23:25:32Z</dc:date>
    </item>
    <item>
      <title>Re: how to repeatedly run regression for larger sets</title>
      <link>https://communities.sas.com/t5/SAS-Programming/how-to-repeatedly-run-regression-for-larger-sets/m-p/78887#M17049</link>
      <description>&lt;HTML&gt;&lt;HEAD&gt;&lt;/HEAD&gt;&lt;BODY&gt;&lt;P&gt;I am no economist or financial expert but the following simulation should warn you of a potential problem when looking for linear relationships between autocorrelated time series :&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;/* Generate 100 simulated daily price series */&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;data test;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;format date yymmdd10. price 7.2;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;call streaminit(658487);&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;do stock = 1 to 100;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; price = 50 + 100 * rand('UNIFORM');&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; do date = '01jan2000'd to '31dec2004'd;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; price = price * (1 + rand("NORMAL", 0.0, 0.01));&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; output;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; end;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; end;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;run;&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;proc sort data=test; by date stock; run;&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;/* Make each simulated stock series into a table column */&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;proc transpose data=test out=table prefix=stock_;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;by date;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;id stock;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;var price;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;run;&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;/* Plot the first three series */&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;title "Three simulated price series";&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;ods graphics / antialiasmax=2000;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;proc sgplot data=table;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;series x=date y=stock_1 /lineattrs=(pattern=1);&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;series x=date y=stock_2 /lineattrs=(pattern=1);&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;series x=date y=stock_3 /lineattrs=(pattern=1);&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;xaxis type=time tickvalueformat=yymmdd10.;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;yaxis label="Price";&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;run;&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;/* Find the 5 best correlations for each series */&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;ods listing select none;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;proc corr data=table best=5 Pearson;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;var stock_:;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;ods output PearsonCorr=PC;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;run;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;ods listing select all;&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;options linesize=132;&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;title "Best correlated price series";&lt;/STRONG&gt;&lt;BR /&gt;&lt;STRONG style="font-size: 12pt; font-family: calibri, verdana, arial, sans-serif;"&gt;proc print data=PC(drop=best1 R1 P1) noobs; run;&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;You will notice that far too many relations appear as highly significant among these totally unrelated time series. Looking at the plot should give you a feeling why this happens.&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt; &lt;IMG alt="SGPlot6.png" class="jive-image-thumbnail jive-image" src="https://communities.sas.com/legacyfs/online/3212_SGPlot6.png" width="450" /&gt;&lt;/P&gt;&lt;P&gt;PG&lt;/P&gt;&lt;/BODY&gt;&lt;/HTML&gt;</description>
      <pubDate>Sat, 16 Mar 2013 02:17:35 GMT</pubDate>
      <guid>https://communities.sas.com/t5/SAS-Programming/how-to-repeatedly-run-regression-for-larger-sets/m-p/78887#M17049</guid>
      <dc:creator>PGStats</dc:creator>
      <dc:date>2013-03-16T02:17:35Z</dc:date>
    </item>
    <item>
      <title>Re: how to repeatedly run regression for larger sets</title>
      <link>https://communities.sas.com/t5/SAS-Programming/how-to-repeatedly-run-regression-for-larger-sets/m-p/78888#M17050</link>
      <description>&lt;HTML&gt;&lt;HEAD&gt;&lt;/HEAD&gt;&lt;BODY&gt;&lt;P&gt;Two ways of looping in SAS - BY statements or macro.&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;You could modify the dataset so that you only have two variables and using a by statement, but a macro seems simpler.&lt;/P&gt;&lt;P&gt;Or you can use Matlab or Octave with your for loops. &lt;/P&gt;&lt;/BODY&gt;&lt;/HTML&gt;</description>
      <pubDate>Sat, 16 Mar 2013 02:43:51 GMT</pubDate>
      <guid>https://communities.sas.com/t5/SAS-Programming/how-to-repeatedly-run-regression-for-larger-sets/m-p/78888#M17050</guid>
      <dc:creator>Reeza</dc:creator>
      <dc:date>2013-03-16T02:43:51Z</dc:date>
    </item>
  </channel>
</rss>

