There are many regression procedures in SAS. I think these three are primers.

1) proc reg     --- general one
2) proc glm    --- easily handle large number of category variables w/o additional coding  
3) mixed  ---- take care of panel data with different types of covariance structures

logistic可以吗？

我看到文献中还有很多其它方法 比如
bayesian
ridge
等等