[Limdep Nlogit List] Conditional (posterior) class probabilities for panel ordered latent class model
Richard Yao
rickyyao at gmail.com
Wed Apr 21 08:15:47 AEST 2021
Dear Everyone,
Greetings from New Zealand!
I am currently running panel ordered latent logit class models on NLOGIT 6
and would like to calculate the conditional (posterior) class
probabilities.
I have used the NLOGIT code below and was able to estimate the panel latent
class model but unfortunately, I could not find the conditional (posterior)
class probabilities in the "CLASSP_I" MATRIX.
I would highly appreciate any thoughts/advice.
Best wishes,
Richard Yao, PhD
SCION, Private Bag 3020
Rotorua 3046, New Zealand
P: 64 7 343 5747; M: 64 21 0247 3234
SAMPLE ; All$
SETPANEL ; Group = pid ; Pds = csnum $
OLOGIT
; Lhs = rank_inv
; Rhs = one, broad_ar, targ_ar, targ_grd, remo_inf, ster_ins, effect
; list
; Partial effects $
OLOGIT
; Lhs = rank_inv
; Rhs = one, broad_ar, targ_ar, targ_grd, remo_inf, ster_ins, effect
; Partial effects
; LCM
; Pts = 3
; Parameters (Save posterior results)
; Panel $
|-> OLOGIT
; Lhs = rank_inv
; Rhs = one, broad_ar, targ_ar, targ_grd, remo_inf, ster_ins, effect
; Partial effects
; LCM
; Pts = 3
; Parameters (Save posterior results)
; List
; Panel $
Line search at iteration 68 does not improve the function
Exiting optimization
-----------------------------------------------------------------------------
Latent Class / Panel OrdProbs Model
Dependent variable RANK_INV
Log likelihood function -77528.46129
Estimation based on N = 63984, K = 35
Inf.Cr.AIC = 155126.9 AIC/N = 2.424
Unbalanced panel has 1333 individuals
Latent class model with 3 latent classes
Ordered probability model
Ordered LOGIT probability model
LHS variable = values 0,1,..., 5
--------+--------------------------------------------------------------------
| Standard Prob. 95% Confidence
RANK_INV| Coefficient Error z |z|>Z* Interval
--------+--------------------------------------------------------------------
|Model parameters for latent class 1.............................
Constant| -3.52320*** .23606 -14.93 .0000 -3.98586 -3.06053
BROAD_AR| 5.95097*** .12678 46.94 .0000 5.70248 6.19946
TARG_AR| 6.92984*** .11222 61.75 .0000 6.70989 7.14979
TARG_GRD| 7.33345*** .10946 67.00 .0000 7.11891 7.54799
REMO_INF| 6.63923*** .11302 58.74 .0000 6.41772 6.86075
STER_INS| 6.33949*** .12028 52.71 .0000 6.10375 6.57523
EFFECT| 5.53569*** .09779 56.61 .0000 5.34403 5.72734
Mu(01)| 3.99266*** .23141 17.25 .0000 3.53910 4.44622
Mu(02)| 5.43495*** .17091 31.80 .0000 5.09997 5.76994
Mu(03)| 6.73958*** .20804 32.40 .0000 6.33183 7.14734
Mu(04)| 8.21072*** .21450 38.28 .0000 7.79031 8.63114
|Model parameters for latent class 2.............................
Constant| -2.52192*** .29340 -8.60 .0000 -3.09697 -1.94688
BROAD_AR| 3.26588*** .08805 37.09 .0000 3.09332 3.43845
TARG_AR| 7.12458*** .10202 69.84 .0000 6.92463 7.32453
TARG_GRD| 8.48640*** .09783 86.74 .0000 8.29465 8.67815
REMO_INF| 8.53274*** .09303 91.72 .0000 8.35039 8.71508
STER_INS| 8.39927*** .09673 86.83 .0000 8.20968 8.58886
EFFECT| 4.11553*** .10494 39.22 .0000 3.90985 4.32122
Mu(01)| 4.38768*** .31560 13.90 .0000 3.76912 5.00624
Mu(02)| 6.49755*** .24170 26.88 .0000 6.02382 6.97128
Mu(03)| 7.90108*** .28617 27.61 .0000 7.34020 8.46195
Mu(04)| 9.34575*** .33196 28.15 .0000 8.69512 9.99638
|Model parameters for latent class 3.............................
Constant| 1.08374*** .33059 3.28 .0010 .43581 1.73168
BROAD_AR| -1.86830*** .06969 -26.81 .0000 -2.00489 -1.73171
TARG_AR| .13114** .06415 2.04 .0409 .00541 .25687
TARG_GRD| 1.16636*** .06240 18.69 .0000 1.04407 1.28866
REMO_INF| 1.81549*** .06787 26.75 .0000 1.68247 1.94851
STER_INS| 1.05045*** .05081 20.67 .0000 .95086 1.15005
EFFECT| 1.41528*** .08621 16.42 .0000 1.24631 1.58425
Mu(01)| 1.24356*** .38216 3.25 .0011 .49453 1.99259
Mu(02)| 2.20918*** .18116 12.19 .0000 1.85412 2.56424
Mu(03)| 3.12403*** .38240 8.17 .0000 2.37453 3.87352
Mu(04)| 4.21616*** .39565 10.66 .0000 3.44069 4.99163
|Estimated prior probabilities for class membership..............
Class1Pr| .61720*** .01381 44.68 .0000 .59012 .64427
Class2Pr| .22585*** .01209 18.67 .0000 .20214 .24955
Class3Pr| .15696*** .01012 15.51 .0000 .13713 .17679
--------+--------------------------------------------------------------------
***, **, * ==> Significance at 1%, 5%, 10% level.
Model was estimated on Apr 21, 2021 at 10:07:48 AM
-----------------------------------------------------------------------------
RP cannot compute ME for ORDERED when J > 9..
More information about the Limdep
mailing list