proc stepwise ;
model y=x1-x3/sle=0.2 sls=0.2;*这里定的;
run;
///////////////////
这样才不会被剔出去。结果不粘了。
反正你看不出,格式是乱的。作者: shiyiming 时间: 2004-11-30 03:16 标题: to ahuige 您干脆单开一个“SAS与统计应用”的板块算了。也省得大家发短信麻烦。众望所归呀作者: shiyiming 时间: 2004-11-30 06:41 标题: re You will have to increase the significant level like what ahuige did. Since both p-values are >0.18 you'd better set the level to 0.2 to let the variable into the model, otherwise both of them will be discarded.作者: shiyiming 时间: 2004-11-30 12:20 标题: re 这个回归这么差,干吗还勉强拟合呢。