quantreg -- Quantile Regression (Version 2.0-2)

This is a submission to compute regression quantiles and some related
rank statistics.  It is a revision of a prior submission to incorporate
some recent develoopments in rank based inference which employ
regression quantile ideas.  The submission  consists of:

	rq.r--A ratfor subroutine 

and several S functions:

	rq--which directly calls rq.o
	trq--which computes analogues of the trimmed mean for regression
	trq.print--which prints ls.print style output for trq
	qrq--which is called by trq
	ranks--which computes a vector of rank statistics from rq() output
	rrs.test--which computes tests of linear hypotheses
	rq.omega--estimates a scale parameter for the covariance matrix
	dn--computes bandwidth for sparsity estimate

These eight functions all reside in a file called rq.s which could be read 
by S with the source() command.  Help files for the S functions are also 
provided in the directory doc and could be moved to the appropriate .Data/.Help 
directory.  

The files were packaged as a unix shar archive.  A Makefile generated by
the S CHAPTER command has been included to automate this process.  It has
been tested on Sun hardware, but may need editing.  With luck you might
try 
		make install
		make load
		make clean

If this works smoothly then invoking Splus in this directory will
automatically load the new rq function and you can proceed from there.

      This software has been submitted to Statlib and may be freely
      used and redistributed for non-commercial purposes.  No guarantees
      are offered or implied.  Comments, bug reports, etc are welcome
      and should be sent to roger@ysidro.econ.uiuc.edu or to
 
 				Roger Koenker
 				Department of Economics
 				University of Illinois
 				Champaign, Illinois, 61820
 
Acknowledgements:  Thanks to all those who have contributed comments
on this software following its initial submission in 1991,  particularly
to Gib Bassett, Steve Portnoy, Jana Jureckova, and Cornelius Gutenbrunner.
Special thanks  to Pin Ng (U. of Houston) for help in preparing the new 
version.
Quantreg -- Quantile Regression (Version 2.0)

This is a submission to compute regression quantiles and some related
rank statistics.  It is a revision of a prior submission to incorporate
some recent develoopments in rank based inference which employ
regression quantile ideas.  The submission  consists of:

	rq.r--A ratfor subroutine 

and several S functions:

	rq--which directly calls rq.o
	trq--which computes analogues of the trimmed mean for regression
	trq.print--which prints ls.print style output for trq
	qrq--which is called by trq
	ranks--which computes a vector of rank statistics from rq() output
	rrs.test--which computes tests of linear hypotheses
	rq.omega--estimates a scale parameter for the covariance matrix
	dn--computes bandwidth for sparsity estimate

These eight functions all reside in a file called rq.s which could be read 
by S with the source() command.  Help files for the S functions are also 
provided in the directory doc and could be moved to the appropriate .Data/.Help 
directory.  

The files were packaged as a unix shar archive.  A Makefile generated by
the S CHAPTER command has been included to automate this process.  It has
been tested on Sun hardware, but may need editing.  With luck you might
try 
		make install
		make load
		make clean

If this works smoothly then invoking Splus in this directory will
automatically load the new rq function and you can proceed from there.

      This software has been submitted to Statlib and may be freely
      used and redistributed for non-commercial purposes.  No guarantees
      are offered or implied.  Comments, bug reports, etc are welcome
      and should be sent to roger@ysidro.econ.uiuc.edu or to
 
 				Roger Koenker
 				Department of Economics
 				University of Illinois
 				Champaign, Illinois, 61820
 
Acknowledgements:  Thanks to all those who have contributed comments
on this software following its initial submission in 1991,  particularly
to Gib Bassett, Steve Portnoy, Jana Jureckova, and Cornelius Gutenbrunner.
Special thanks  to Pin Ng (U. of Houston) for help in preparing the new 
version.
Quantreg -- Quantile Regression (Version 2.0)

This is a submission to compute regression quantiles and some related
rank statistics.  It is a revision of a prior submission to incorporate
some recent develoopments in rank based inference which employ
regression quantile ideas.  The submission  consists of:

	rq.r--A ratfor subroutine 

and several S functions:

	rq--which directly calls rq.o
	trq--which computes analogues of the trimmed mean for regression
	trq.print--which prints ls.print style output for trq
	qrq--which is called by trq
	ranks--which computes a vector of rank statistics from rq() output
	rrs.test--which computes tests of linear hypotheses
	rq.omega--estimates a scale parameter for the covariance matrix
	dn--computes bandwidth for sparsity estimate

These eight functions all reside in a file called rq.s which could be read 
by S with the source() command.  Help files for the S functions are also 
provided in the directory doc and could be moved to the appropriate .Data/.Help 
directory.  

The files were packaged as a unix shar archive.  A Makefile generated by
the S CHAPTER command has been included to automate this process.  It has
been tested on Sun hardware, but may need editing.  With luck you might
try 
		make install
		make load
		make clean

If this works smoothly then invoking Splus in this directory will
automatically load the new rq function and you can proceed from there.

      This software has been submitted to Statlib and may be freely
      used and redistributed for non-commercial purposes.  No guarantees
      are offered or implied.  Comments, bug reports, etc are welcome
      and should be sent to roger@ysidro.econ.uiuc.edu or to
 
 				Roger Koenker
 				Department of Economics
 				University of Illinois
 				Champaign, Illinois, 61820
 
Acknowledgements:  Thanks to all those who have contributed comments
on this software following its initial submission in 1991,  particularly
to Gib Bassett, Steve Portnoy, Jana Jureckova, and Cornelius Gutenbrunner.
Special thanks  to Pin Ng (U. of Houston) for help in preparing the new 
version.