[Limdep Nlogit List] Latent class model converged after a huge number of iterations
William Greene
wgreene at stern.nyu.edu
Tue Jun 30 23:48:37 AEST 2020
numbers like that suggest that the model is not identified. The
interaction of
the variables is the usual culprit. In a nonlinear model, with latent
classes, it
will be more complicated than multicollinearity that you can see directly.
I would
suggest you fit a much smaller model and build up the model you want slowly.
The misbehavior of the specification will show up at some point along the
way.
/B. Greene
On Mon, Jun 29, 2020 at 8:32 PM Thao Thai via Limdep <
limdep at mailman.sydney.edu.au> wrote:
> Hi Nlogit users,
>
> I am running a latent class model with 2 classes. The model converged after
> 177 iterations but some coefficients are huge with p-values close to 1
> (e.g. camt50, pamt50,etc.).
> Could you please kindly let me know why the model reports such strange
> results even thought the model converged?
>
> My 2 class LC results are below. Any suggestions would be greatly
> appreciated.
> Thank you so much.
> Thao
>
> -----------------------------------------------------------------------------
> Discrete choice (multinomial logit) model
> Dependent variable Choice
> Log likelihood function -3027.91231
> Estimation based on N = 1881, K = 51
> Inf.Cr.AIC = 6157.8 AIC/N = 3.274
> ---------------------------------------
> Log likelihood R-sqrd R2Adj
> ASCs only model must be fit separately
> Use NLOGIT ;...;RHS=ONE$
> Note: R-sqrd = 1 - logL/Logl(constants)
> ---------------------------------------
> Chi-squared[46] = 634.49410
> Prob [ chi squared > value ] = .00000
> Response data are given as ind. choices
> Number of obs.= 2434, skipped 553 obs
>
> --------+--------------------------------------------------------------------
> | Standard Prob. 95% Confidence
> CHO| Coefficient Error z |z|>Z* Interval
>
> --------+--------------------------------------------------------------------
> RL_H1|1| .07880 .18947 .42 .6775 -.29256 .45015
> RL_H2|1| .32233** .15780 2.04 .0411 .01304 .63161
> FL_|1| .15026** .06411 2.34 .0191 .02461 .27591
> CR_H1|1| .30369** .15393 1.97 .0485 .00198 .60539
> CR_2|1| .07018 .10914 .64 .5203 -.14374 .28409
> LO_H1|1| -.48240*** .14152 -3.41 .0007 -.75978 -.20502
> SA_|1| .01294*** .00098 13.18 .0000 .01102 .01487
> HISHOS|1| .53228*** .13065 4.07 .0000 .27621 .78834
> COM|1| -.06172 .39015 -.16 .8743 -.82640 .70296
> RL_C1|1| .33323* .19321 1.72 .0846 -.04546 .71193
> RL_C2|1| .19311 .17713 1.09 .2756 -.15406 .54029
> CR_C1|1| .14782 .15909 .93 .3528 -.16399 .45963
> LO_C1|1| -.33950** .15086 -2.25 .0244 -.63519 -.04382
> LO_C2|1| -.79541*** .17740 -4.48 .0000 -1.14311 -.44771
> CFEMAL|1| -.50458*** .17382 -2.90 .0037 -.84525 -.16391
> CA3549|1| .21543 .19597 1.10 .2716 -.16867 .59953
> CAMT50|1| 1.05281*** .20358 5.17 .0000 .65380 1.45183
> HISCOM|1| .35769 .30701 1.17 .2440 -.24403 .95941
> PRI|1| .72894*** .23093 3.16 .0016 .27633 1.18155
> RL_P1|1| .01445 .14051 .10 .9181 -.26095 .28984
> CR_P1|1| .28539* .14898 1.92 .0554 -.00660 .57739
> LO_P1|1| -1.03715*** .16408 -6.32 .0000 -1.35875 -.71556
> LO_P2|1| -.86538*** .15753 -5.49 .0000 -1.17412 -.55663
> PFEMAL|1| -.01737 .17614 -.10 .9215 -.36260 .32787
> PA3549|1| -.12983 .16126 -.81 .4207 -.44590 .18623
> PAMT50|1| .21489 .18085 1.19 .2347 -.13956 .56934
> HISPRI|1| .37596** .16405 2.29 .0219 .05443 .69750
> IND|1| -.44619* .23633 -1.89 .0590 -.90938 .01700
> RL_I1|1| .50658*** .16963 2.99 .0028 .17411 .83906
> RL_I2|1| .85985*** .17568 4.89 .0000 .51552 1.20417
> CR_I1|1| .51109*** .13898 3.68 .0002 .23871 .78348
> LO_I1|1| -.55538*** .14041 -3.96 .0001 -.83058 -.28018
> IFEMAL|1| -.61910*** .17288 -3.58 .0003 -.95794 -.28026
> HISIND|1| .96531*** .17646 5.47 .0000 .61945 1.31116
> GOV|1| .38918* .23520 1.65 .0980 -.07181 .85017
> RL_G1|1| -.35640** .14665 -2.43 .0151 -.64382 -.06897
> CR_G1|1| .43164*** .13869 3.11 .0019 .15981 .70347
> LO_G1|1| -.57675*** .14093 -4.09 .0000 -.85297 -.30052
> GFEMAL|1| -.33687* .17868 -1.89 .0594 -.68707 .01332
> GA3549|1| -.03689 .19169 -.19 .8474 -.41259 .33882
> GAMT50|1| .20023 .21447 .93 .3505 -.22012 .62057
> HISGOV|1| .60211*** .14132 4.26 .0000 .32513 .87908
> NON|1| .12186 .24794 .49 .6231 -.36408 .60781
> RL_N1|1| -.00469 .15976 -.03 .9766 -.31781 .30844
> CR_N1|1| .36382** .16054 2.27 .0234 .04917 .67846
> LO_N1|1| -.68641*** .18403 -3.73 .0002 -1.04710 -.32572
> LO_N2|1| -.44462** .18561 -2.40 .0166 -.80840 -.08083
> NFEMAL|1| -.32362* .19138 -1.69 .0908 -.69871 .05147
> NA3549|1| -.03864 .20620 -.19 .8513 -.44278 .36549
> NAMT50|1| .21775 .22891 .95 .3415 -.23091 .66641
> HISNON|1| .39563** .15520 2.55 .0108 .09145 .69982
>
> --------+--------------------------------------------------------------------
> ***, **, * ==> Significance at 1%, 5%, 10% level.
> Model was estimated on Jun 30, 2020 at 09:47:50 AM
>
> -----------------------------------------------------------------------------
>
> Iterative procedure has converged
> Normal exit: 177 iterations. Status=0, F= .2966264D+04
>
>
> -----------------------------------------------------------------------------
> Latent Class Logit Model
> Dependent variable CHO
> Log likelihood function -2966.26364
> Restricted log likelihood -3370.29956
> Chi squared [103](P= .000) 808.07183
> Significance level .00000
> McFadden Pseudo R-squared .1198813
> Estimation based on N = 1881, K = 103
> Inf.Cr.AIC = 6138.5 AIC/N = 3.263
> ---------------------------------------
> Log likelihood R-sqrd R2Adj
> No coefficients -3370.2996 .1199 .1101
> Constants only can be computed directly
> Use NLOGIT ;...;RHS=ONE$
> At start values -3027.9417 .0204 .0095
> Note: R-sqrd = 1 - logL/Logl(constants)
> ---------------------------------------
> Response data are given as ind. choices
> Number of latent classes = 2
> Average Class Probabilities
> .259 .476
> BHHH estimator used for asymp. variance
> Number of obs.= 2434, skipped 553 obs
>
> --------+--------------------------------------------------------------------
> | Standard Prob. 95% Confidence
> CHO| Coefficient Error z |z|>Z* Interval
>
> --------+--------------------------------------------------------------------
> |Random utility parameters in latent class -->>
> 1...................
> RL_H1|1| .57314 .66678 .86 .3900 -.73373 1.88001
> RL_H2|1| -.02754 .52799 -.05 .9584 -1.06238 1.00730
> FL_|1| .13144 .22894 .57 .5659 -.31728 .58015
> CR_H1|1| .73319 .62068 1.18 .2375 -.48332 1.94971
> CR_2|1| -.10967 .42208 -.26 .7950 -.93693 .71760
> LO_H1|1| -.85008 .52985 -1.60 .1086 -1.88857 .18841
> SA_|1| .03102*** .00520 5.96 .0000 .02082 .04121
> HISHOS|1| .29351 .38551 .76 .4464 -.46208 1.04910
> COM|1| -2.02608 1.27137 -1.59 .1110 -4.51792 .46576
> RL_C1|1| .54311 .67716 .80 .4225 -.78411 1.87032
> RL_C2|1| -1.54466** .70163 -2.20 .0277 -2.91982 -.16950
> CR_C1|1| .35293 .57112 .62 .5366 -.76643 1.47230
> LO_C1|1| .19103 .57447 .33 .7395 -.93492 1.31698
> LO_C2|1| .25908 .59437 .44 .6629 -.90587 1.42403
> CFEMAL|1| -2.28025*** .71548 -3.19 .0014 -3.68256 -.87795
> CA3549|1| 1.78554** .89174 2.00 .0453 .03777 3.53331
> CAMT50|1| 46.5436 .8389D+07 .00 1.0000 ***********
> ***********
> HISCOM|1| 1.15176 .81065 1.42 .1554 -.43709 2.74061
> PRI|1| -.01710 .80730 -.02 .9831 -1.59937 1.56517
> RL_P1|1| .35581 .53673 .66 .5074 -.69617 1.40778
> CR_P1|1| .65787 .55078 1.19 .2323 -.42164 1.73738
> LO_P1|1| -2.65965*** .74254 -3.58 .0003 -4.11501 -1.20429
> LO_P2|1| -1.88272*** .61240 -3.07 .0021 -3.08299 -.68244
> PFEMAL|1| -.13393 .61030 -.22 .8263 -1.33010 1.06225
> PA3549|1| .33681 .56773 .59 .5530 -.77591 1.44953
> PAMT50|1| 43.1245 .8389D+07 .00 1.0000 ***********
> ***********
> HISPRI|1| 3.58758*** .79474 4.51 .0000 2.02992 5.14524
> IND|1| -3.53672*** 1.02049 -3.47 .0005 -5.53685 -1.53659
> RL_I1|1| 2.12753*** .77133 2.76 .0058 .61575 3.63932
> RL_I2|1| 3.87220*** .85688 4.52 .0000 2.19274 5.55165
> CR_I1|1| 2.24454*** .58639 3.83 .0001 1.09524 3.39385
> LO_I1|1| -2.95126*** .66040 -4.47 .0000 -4.24562 -1.65689
> IFEMAL|1| -.90993 .58295 -1.56 .1185 -2.05249 .23264
> HISIND|1| 3.36155*** .78707 4.27 .0000 1.81893 4.90418
> GOV|1| 1.26511** .60911 2.08 .0378 .07128 2.45894
> RL_G1|1| -1.38467** .59707 -2.32 .0204 -2.55490 -.21445
> CR_G1|1| -.97856** .47617 -2.06 .0399 -1.91184 -.04529
> LO_G1|1| -1.07453*** .36775 -2.92 .0035 -1.79531 -.35375
> GFEMAL|1| -.75502 .48325 -1.56 .1182 -1.70217 .19213
> GA3549|1| .91479* .48883 1.87 .0613 -.04330 1.87289
> GAMT50|1| 43.4431 .8389D+07 .00 1.0000 ***********
> ***********
> HISGOV|1| 1.17754*** .32933 3.58 .0003 .53206 1.82302
> NON|1| -5.93625** 2.79782 -2.12 .0339 -11.41988 -.45262
> RL_N1|1| -2.62001** 1.17294 -2.23 .0255 -4.91892 -.32110
> CR_N1|1| 7.17512*** 2.53297 2.83 .0046 2.21059 12.13964
> LO_N1|1| -42.4504 .2109D+18 .00 1.0000 ***********
> ***********
> LO_N2|1| -4.40510*** 1.67966 -2.62 .0087 -7.69717 -1.11304
> NFEMAL|1| 1.94069 1.32904 1.46 .1442 -.66418 4.54555
> NA3549|1| -28.9062 .4917D+12 .00 1.0000 ***********
> ***********
> NAMT50|1| 44.0873 .8389D+07 .00 1.0000 ***********
> ***********
> HISNON|1| -.23243 .92295 -.25 .8012 -2.04139 1.57653
> |Random utility parameters in latent class -->>
> 2...................
> RL_H1|2| -.03141 .33713 -.09 .9258 -.69217 .62935
> RL_H2|2| .36069 .29105 1.24 .2152 -.20975 .93113
> FL_|2| .15843 .10226 1.55 .1213 -.04200 .35886
> CR_H1|2| .23102 .27718 .83 .4046 -.31225 .77429
> CR_2|2| .12892 .17411 .74 .4590 -.21234 .47018
> LO_H1|2| -.35813 .24251 -1.48 .1397 -.83344 .11719
> SA_|2| .00812*** .00177 4.59 .0000 .00466 .01159
> HISHOS|2| .75600*** .24425 3.10 .0020 .27728 1.23472
> COM|2| .19671 .72848 .27 .7871 -1.23108 1.62450
> RL_C1|2| .27931 .33839 .83 .4091 -.38393 .94255
> RL_C2|2| .73297** .30638 2.39 .0167 .13247 1.33346
> CR_C1|2| .22909 .26220 .87 .3823 -.28482 .74300
> LO_C1|2| -.57897** .26006 -2.23 .0260 -1.08868 -.06925
> LO_C2|2| -1.19701*** .30833 -3.88 .0001 -1.80132 -.59270
> CFEMAL|2| .19462 .33875 .57 .5656 -.46932 .85855
> CA3549|2| -.34064 .41576 -.82 .4126 -1.15552 .47424
> CAMT50|2| -.45795 .49982 -.92 .3595 -1.43757 .52167
> HISCOM|2| .43151 .48333 .89 .3720 -.51580 1.37882
> PRI|2| 1.06274** .52110 2.04 .0414 .04141 2.08408
> RL_P1|2| -.10678 .23742 -.45 .6529 -.57211 .35856
> CR_P1|2| .01438 .29670 .05 .9613 -.56714 .59590
> LO_P1|2| -.39739 .28311 -1.40 .1604 -.95227 .15748
> LO_P2|2| -.40739 .29272 -1.39 .1640 -.98111 .16634
> PFEMAL|2| .07543 .29394 .26 .7975 -.50068 .65155
> PA3549|2| -.42642 .39242 -1.09 .2772 -1.19554 .34270
> PAMT50|2| -.37411 .39076 -.96 .3384 -1.13998 .39177
> HISPRI|2| -42.2734 .5483D+18 .00 1.0000 ***********
> ***********
> IND|2| .83708 .52394 1.60 .1101 -.18982 1.86398
> RL_I1|2| .05190 .28791 .18 .8570 -.51239 .61619
> RL_I2|2| -.62410 .46110 -1.35 .1759 -1.52784 .27963
> CR_I1|2| -.47968 .36629 -1.31 .1903 -1.19760 .23824
> LO_I1|2| .54665 .35751 1.53 .1263 -.15407 1.24736
> IFEMAL|2| -.67148** .30608 -2.19 .0282 -1.27139 -.07158
> HISIND|2| .00721 .53506 .01 .9892 -1.04148 1.05590
> GOV|2| -.88733 .71577 -1.24 .2151 -2.29022 .51557
> RL_G1|2| 1.28792*** .45320 2.84 .0045 .39967 2.17617
> CR_G1|2| 1.82566*** .49813 3.67 .0002 .84934 2.80197
> LO_G1|2| -1.10454*** .39979 -2.76 .0057 -1.88811 -.32097
> GFEMAL|2| -.26627 .36830 -.72 .4697 -.98812 .45557
> GA3549|2| -1.05627** .53076 -1.99 .0466 -2.09653 -.01601
> GAMT50|2| -.84044 .51774 -1.62 .1045 -1.85519 .17430
> HISGOV|2| .36972 .33570 1.10 .2707 -.28824 1.02769
> NON|2| 1.18387** .48948 2.42 .0156 .22451 2.14324
> RL_N1|2| -.14201 .20773 -.68 .4942 -.54916 .26514
> CR_N1|2| -.27945 .23804 -1.17 .2404 -.74601 .18710
> LO_N1|2| -.24170 .25171 -.96 .3369 -.73505 .25165
> LO_N2|2| -.38703 .24601 -1.57 .1157 -.86919 .09514
> NFEMAL|2| -.35760 .26831 -1.33 .1826 -.88348 .16827
> NA3549|2| -.22911 .38392 -.60 .5507 -.98157 .52336
> NAMT50|2| -.74281* .41928 -1.77 .0765 -1.56458 .07896
> HISNON|2| .39006** .19277 2.02 .0430 .01224 .76787
> |Estimated latent class
> probabilities................................
> PrbCls1| .42517*** .04852 8.76 .0000 .33007 .52028
> PrbCls2| .57483*** .04852 11.85 .0000 .47972 .66993
>
> --------+--------------------------------------------------------------------
> nnnnn.D-xx or D+xx => multiply by 10 to -xx or +xx.
> ***, **, * ==> Significance at 1%, 5%, 10% level.
> Model was estimated on Jun 30, 2020 at 09:49:57 AM
>
> -----------------------------------------------------------------------------
> _______________________________________________
> Limdep site list
> Limdep at mailman.sydney.edu.au
> http://limdep.itls.usyd.edu.au
>
>
--
William Greene
Department of Economics, emeritus
Stern School of Business, New York University
44 West 4 St.
New York, NY, 10012
URL: https://protect-au.mimecast.com/s/5dRXCJyBrGf59nowFVej-d?domain=people.stern.nyu.edu
Email: wgreene at stern.nyu.edu
Ph. +1.646.596.3296
Editor in Chief: Journal of Productivity Analysis
Editor in Chief: Foundations and Trends in Econometrics
Associate Editor: Economics Letters
Associate Editor: Journal of Business and Economic Statistics
More information about the Limdep
mailing list