自己顶一下，请高手帮忙！

 Quantile regression extends that technique to estimate conditional quantiles of the distribution of y.  Median regression is the 50th quantile, but you can specify any quantiles you want.  If you specify more than one, you can get a plot of the coefficients against the quantiles as well as the estimates for each quantile.  Thus quantile regression allows you to estimate the y distribution more flexibly than ordinary regression, and it can be helpful in the face of outliers and violations of the usual least squares assumptions about the error distribution.  You can find a discussion of some of the mathematical, computational, and practical issues at everybody’s favorite statistics site :-) 

 

SPSSINC QUANTREG is implemented through the R quantreg package.  Like all of our R extension packages, it requires R 2.7, the R plug-in, and the Python plug-in.  Specifically for quantreg, you need to install the R Quantreg package, and the SPSSINC QUANTREG extension package.  (To install an R package, start R and use the Packages>Install Packages menu item.)  The procedure includes a dialog box interface as well as the command syntax definition, and the output is SPSS pivot tables and, optionally, a set of plots.  Including graphical output from R in the Viewer is a new feature of SPSS Statistics 17.  The principal author of the R package is Roger Koenker, who is the best known researcher in this area.

分位数回归不是用SPSS做的  用r 或state

55555555555555555555555

spss不能做 eviews可以