用Tobit 模型做了回歸,現(xiàn)有兩個(gè)方案,但因?qū)W業(yè)不專,難以做出選擇,望有高手指點(diǎn)
Coefficient Std. Error z-Statistic Prob.
C 1.676197 0.263858 6.352642 0.0000
D1T 0.080134 0.037324 2.146974 0.0318
DEPO -5.76E-06 1.83E-06 -3.143190 0.0017
INCOME -1.75E-05 3.94E-06 -4.430505 0.0000
AGE -0.765601 0.394130 -1.942509 0.0521
MINO 0.040797 0.033592 1.214489 0.2246
Error Distribution
SCALE:C(7) 0.062561 0.007945 7.874068 0.0000
R-squared 0.801610 Mean dependent var 0.713806
Adjusted R-squared 0.752013 S.D. dependent var 0.142813
S.E. of regression 0.071119 Akaike info criterion -2.253729
Sum squared resid 0.121389 Schwarz criterion -1.929925
Log likelihood 41.93280 Hannan-Quinn criter. -2.148177
Avg. log likelihood 1.352671
Left censored obs 0 Right censored obs 0
Uncensored obs 31 Total obs 31
另一個(gè)是
Coefficient Std. Error z-Statistic Prob.
C 1.630011 0.296630 5.495107 0.0000
AGE -0.845922 0.444848 -1.901600 0.0572
INCOME -9.46E-06 4.61E-06 -2.051628 0.0402
DEPO -6.91E-06 1.96E-06 -3.522884 0.0004
KNOW -0.718256 0.432826 -1.659456 0.0970
TRAN 0.010183 0.006274 1.623192 0.1045
Error Distribution
SCALE:C(7) 0.064341 0.008171 7.874068 0.0000
R-squared 0.789930 Mean dependent var 0.713806
Adjusted R-squared 0.737412 S.D. dependent var 0.142813
S.E. of regression 0.073182 Akaike info criterion -2.197624
Sum squared resid 0.128536 Schwarz criterion -1.873820
Log likelihood 41.06317 Hannan-Quinn criter. -2.092072
Avg. log likelihood 1.324618
Left censored obs 0 Right censored obs 0
Uncensored obs 31 Total obs 31
格式比較混亂