{smcl}
{* 15 Feb 2007}{...}
{cmd:help plssas}
{hline}

{title:Title}

    {hi: SAS's PROC PLS (Partial Least Squares)}

{title:Syntax}

{p 8 17 2}
{cmdab:plssas}
[varlist]
[{cmd:,} {it:options}]

{synoptset 20 tabbed}{...}
{synopthdr}
{synoptline}
{syntab:Main}
{synopt:{opt y(varlist)}} specify a list of y-variables.{p_end}
{synopt:{opt x(varlist)}} specify a list of x-variables.{p_end}
{synopt:{opt n:fac(integer 0)}} specify the number of factors to extract.{p_end}
{synopt:{opt nosas}} specify that SAS should not be run (only used when the output files exist).{p_end}
{synopt:{opt nofile}} specifies that the Stata dataset is not exported to a csv file.{p_end}
{synopt:{opt m:ethod(string)}} specify the general factor extraction method to use, the default is PLS partial least squares.{p_end}
{synopt:{opt p:refix(string)}} specify a prefix to be added to the names of the output csv files.{p_end}
{synopt:{opt exe(string asis)}} specify the path and executable to run SAS.{p_end}
{synoptline}
{p2colreset}{...}

{title:Description}

{pstd}
{cmd:plssas} creates a *.sas program to run a PLS analysis, then runs this file in the background and the output datasets created by SAS 
are converted to *.CSV files. The set of files of OUTPUT csv files are:

      yweights.csv
      xweights.csv
      xload.csv
      xeff.csv
      pest.csv
      perc.csv
      csp.csv
      codedcoef.csv
      out.csv

{pstd}
Also a file called temp.sas will contain the syntax and temp.csv will be the dataset created ready for importing into SAS.

{pstd} The default version of SAS used is 9.1 and the path is by default {bf: c:\Program Files\SAS\SAS 9.1\sas.exe}.

{pstd} A full description of the PLS procedure can be found in the SAS help files. Not all the options are implemented but might be added on
request.

{pstd}
A SAS logfile will also be created called temp.log. 

{title:Options}

{dlgtab:Main}

{phang}
{opt y(varlist)} specifies the list of y-variables or responses.

{phang}
{opt x(varlist)} specifies the list of x-variables or predictors.

{phang}
{opt n:fac(integer 0)} specifies the number of factors to extract. The default is exactly the same as the SAS default of the minimum of
15 and the number of predictors.

{phang}
{opt nosas} specifies that SAS will not be executed. This option can only be used if the output csv files from a previous analysis exist in the current
directory.

{phang}
{opt nofile} specifies that the Stata dataset is not exported to a csv file. This option should only be used when the output files already exist
and hence this command is being used to replay results from a previous analysis.

{phang}
{opt m:ethod(string)} specifies the general factor extraction method to use. The default is PLS partial least squares, SIMPLS uses the method of de Jong (1993), 
PCR is principal components regression, and RRR is reduced rank regression.

{phang}
{opt p:refix(string)} specifies a prefix to the csv output filenames. 

{phang}
{opt exe(string asis)} specifies the path and name of the SAS executable; note spaces are fine there is no need to use quotes in this option.
 At present this has been implemented for only the Windows version but 
will in the future be extended to both Unix and Mac operating systems. The default is the path {bf: c:\Program Files\SAS\SAS 9.1\sas.exe}.

{title:Examples}

{phang}
{input: plssas, y(fibd99 fat99) x(fber995)}

{phang}
{input: plssas, y(fibd99 fat99) x(fber995)}

{phang}
{input: plssas, y(fibd99 fat99) x(fber995) exe(P:\sas_network_image_8\SAS\sas.exe)}

{title:Author}

Adrian Mander, MRC Biostatistics Unit Hub for Trials Methodology Research, Cambridge, UK.

Email {browse "mailto:adrian.mander@mrc-hnr.cam.ac.uk":adrian.mander@mrc-bsu.cam.ac.uk}

{title:See Also}

Related commands:

{help mvreg}