SAS LOGISTIC 逻辑回归中加(EVENT='1')和不加(EVENT='1')区别

区别在于：最大似然估计分析中估计是刚好正负对调
加上EVENT：
%LET DVVAR = Y;
%LET LOGIT_IN = S.T3;
%LET LOGIT_MODEL = S.Model_Params;
%LET LOGIT_SCORE = S.Pred_Probs;

%let VarList= X1_WOE--B&BN._WOE;

/* Storing the results of the model in a dataset */
proc logistic data=&LOGIT_IN OUTEST=&LOGIT_MODEL;
model &DVVAR (event='1')= &VarList /
selection =stepwise sls=0.05 sle=0.05;
OUTPUT OUT=&LOGIT_SCORE P=Pred_Y;
run;

输出结果
最大似然估计分析

标准 Wald
参数 自由度 估计 误差 卡方 Pr > 卡方

Intercept 1 -0.2769 0.0618 20.0856 <.0001
X1_WOE 1 0.8903 0.2490 12.7851 0.0003
X3_WOE 1 1.0583 0.1558 46.1674 <.0001
X4_WOE 1 1.0319 0.1264 66.6874 <.0001
B1_WOE 1 0.8293 0.4066 4.1600 0.0414

没有加上EVENT：
%LET DVVAR = Y;
%LET LOGIT_IN = S.T3;
%LET LOGIT_MODEL = S.Model_Params;
%LET LOGIT_SCORE = S.Pred_Probs;

%let VarList= X1_WOE--B&BN._WOE;

/* Storing the results of the model in a dataset */
proc logistic data=&LOGIT_IN OUTEST=&LOGIT_MODEL;
model &DVVAR= &VarList /
selection =stepwise sls=0.05 sle=0.05;
OUTPUT OUT=&LOGIT_SCORE P=Pred_Y;
run;

输出结果：

最大似然估计分析

标准 Wald
参数 自由度 估计 误差 卡方 Pr > 卡方

Intercept 1 0.2769 0.0618 20.0856 <.0001
X1_WOE 1 -0.8903 0.2490 12.7851 0.0003
X3_WOE 1 -1.0583 0.1558 46.1674 <.0001
X4_WOE 1 -1.0319 0.1264 66.6874 <.0001
B1_WOE 1 -0.8293 0.4066 4.1600 0.0414