Important: Read this before posting to this forum

  1. This forum is for questions related to the use of Apollo. We will answer some general choice modelling questions too, where appropriate, and time permitting. We cannot answer questions about how to estimate choice models with other software packages.
  2. There is a very detailed manual for Apollo available at http://www.ApolloChoiceModelling.com/manual.html. This contains detailed descriptions of the various Apollo functions, and numerous examples are available at http://www.ApolloChoiceModelling.com/examples.html. In addition, help files are available for all functions, using e.g. ?apollo_mnl
  3. Before asking a question on the forum, users are kindly requested to follow these steps:
    1. Check that the same issue has not already been addressed in the forum - there is a search tool.
    2. Ensure that the correct syntax has been used. For any function, detailed instructions are available directly in Apollo, e.g. by using ?apollo_mnl for apollo_mnl
    3. Check the frequently asked questions section on the Apollo website, which discusses some common issues/failures. Please see http://www.apollochoicemodelling.com/faq.html
    4. Make sure that R is using the latest official release of Apollo.
  4. If the above steps do not resolve the issue, then users should follow these steps when posting a question:
    1. provide full details on the issue, including the entire code and output, including any error messages
    2. posts will not immediately appear on the forum, but will be checked by a moderator first. This may take a day or two at busy times. There is no need to submit the post multiple times.

Result change drastically when increase halton draws from 1500 to 2000 draws

Ask questions about how to estimate models and how to change your settings for estimation.
Post Reply
sethyash52
Posts: 11
Joined: 09 Sep 2022, 06:38

Result change drastically when increase halton draws from 1500 to 2000 draws

Post by sethyash52 »

Hi Stehane,

I am considering log-normal distribution in all these results. Why am I getting NaN when I increase the Halton draws to 2000? Also, the parameter estimates change drastically when I change the number of Halton draws from 1500 to 2000.

Thanks

Code: Select all

Estimates:
  Estimate        s.e.   t.rat.(0)    Rob.s.e. Rob.t.rat.(0)
mu_log_b_pp          -6.261434    0.679268    -9.21792    0.604042     -10.36589
sigma_log_b_pp       -1.978358    0.413328    -4.78641    0.310212      -6.37743
mu_log_b_oc          -8.808189    0.900963    -9.77641    0.454237     -19.39118
sigma_log_b_oc       -5.986835    0.633005    -9.45780    0.256070     -23.37966
mu_log_b_ct          -9.919012    1.079592    -9.18774    0.487420     -20.35001
sigma_log_b_ct      -10.647421    1.159454    -9.18313    0.462790     -23.00701
b_r                   0.090107    0.027334     3.29648    0.026901       3.34958
b_cf                  0.232221    0.089759     2.58716    0.087898       2.64195
b_e                  -0.166802    0.096987    -1.71983    0.097086      -1.71809
asc_CV                0.000000          NA          NA          NA            NA
asc_EV               -0.209528    0.239563    -0.87462    0.222123      -0.94330
b_Male                0.066565    0.095290     0.69855    0.091213       0.72978
b_female              0.000000          NA          NA          NA            NA
b_ageAvg              0.012343    0.006230     1.98113    0.005867       2.10358
b_dailyDistanceAvg   -0.006702    0.002608    -2.56933    0.002608      -2.56988
b_knowledgeMiddle     0.346267    0.115882     2.98810    0.106528       3.25047
b_knowledgeLow        0.000000          NA          NA          NA            NA
b_knowledgeHigh      -0.003681    0.136485    -0.02697    0.126756      -0.02904
b_priorityLanes       0.173473    0.101335     1.71187    0.098229       1.76601
b_freeParking         0.119394    0.092157     1.29556    0.090466       1.31978
b_tollExemption       0.000000          NA          NA          NA            NA

Number of inter-individual draws            : 200 (halton)

LL(start)                                   : -4479.28
LL at equal shares, LL(0)                   : -3609.22
LL at observed shares, LL(C)                : -3567.91
LL(final)                                   : -3281.55
Rho-squared vs equal shares                  :  0.0908 
Adj.Rho-squared vs equal shares              :  0.0861 
Rho-squared vs observed shares               :  0.0803 
Adj.Rho-squared vs observed shares           :  0.0755 
AIC                                         :  6597.1 
BIC                                         :  6708.58 

####################################################################################################################
Estimates:
  Estimate        s.e.   t.rat.(0)    Rob.s.e. Rob.t.rat.(0)
mu_log_b_pp          -5.896880    0.623216     -9.4620    0.563972      -10.4560
sigma_log_b_pp       -1.538922    0.415741     -3.7016    0.304662       -5.0512
mu_log_b_oc          -9.403876    0.448479    -20.9684    0.328346      -28.6401
sigma_log_b_oc       -6.088848    0.169605    -35.9002    0.033204     -183.3756
mu_log_b_ct          -9.086248    0.397846    -22.8386    0.310191      -29.2924
sigma_log_b_ct      -10.638934    0.219760    -48.4116    0.019384     -548.8423
b_r                   0.088187    0.027173      3.2454    0.026774        3.2938
b_cf                  0.225512    0.089366      2.5235    0.087534        2.5763
b_e                  -0.159770    0.096673     -1.6527    0.096794       -1.6506
asc_CV                0.000000          NA          NA          NA            NA
asc_EV               -0.136107    0.236734     -0.5749    0.221219       -0.6153
b_Male                0.074629    0.094022      0.7937    0.090262        0.8268
b_female              0.000000          NA          NA          NA            NA
b_ageAvg              0.011618    0.006130      1.8952    0.005804        2.0017
b_dailyDistanceAvg   -0.006940    0.002592     -2.6774    0.002587       -2.6827
b_knowledgeMiddle     0.369791    0.114261      3.2364    0.105357        3.5099
b_knowledgeLow        0.000000          NA          NA          NA            NA
b_knowledgeHigh       0.013946    0.135392      0.1030    0.126262        0.1105
b_priorityLanes       0.157775    0.100860      1.5643    0.097877        1.6120
b_freeParking         0.114210    0.091765      1.2446    0.090106        1.2675
b_tollExemption       0.000000          NA          NA          NA            NA

Number of inter-individual draws            : 500 (halton)

LL(start)                                   : -4479.28
LL at equal shares, LL(0)                   : -3609.22
LL at observed shares, LL(C)                : -3567.91
LL(final)                                   : -3281.41
Rho-squared vs equal shares                  :  0.0908 
Adj.Rho-squared vs equal shares              :  0.0861 
Rho-squared vs observed shares               :  0.0803 
Adj.Rho-squared vs observed shares           :  0.0755 
AIC                                         :  6596.82 
BIC                                         :  6708.3 

########################################################################################################################
Estimates:
  Estimate        s.e.   t.rat.(0)    Rob.s.e. Rob.t.rat.(0)
mu_log_b_pp          -5.825467    0.422017   -13.80388    0.405565     -14.36382
sigma_log_b_pp        1.484376    0.172701     8.59505    0.133522      11.11705
mu_log_b_oc          -8.214120    0.443839   -18.50696    0.316805     -25.92803
sigma_log_b_oc       -5.256324    0.261759   -20.08078    0.076134     -69.04001
mu_log_b_ct          -8.355631    0.633229   -13.19528    0.349301     -23.92099
sigma_log_b_ct       -9.647314    0.594411   -16.23004    0.168365     -57.30014
b_r                   0.089078    0.027220     3.27252    0.026803       3.32340
b_cf                  0.228006    0.089512     2.54722    0.087669       2.60076
b_e                  -0.159974    0.096738    -1.65367    0.096943      -1.65018
asc_CV                0.000000          NA          NA          NA            NA
asc_EV               -0.155333    0.237896    -0.65295    0.222179      -0.69914
b_Male                0.070932    0.094569     0.75006    0.090767       0.78147
b_female              0.000000          NA          NA          NA            NA
b_ageAvg              0.012040    0.006161     1.95416    0.005832       2.06451
b_dailyDistanceAvg   -0.006951    0.002606    -2.66686    0.002605      -2.66878
b_knowledgeMiddle     0.364069    0.114937     3.16754    0.105905       3.43769
b_knowledgeLow        0.000000          NA          NA          NA            NA
b_knowledgeHigh       0.011730    0.136050     0.08622    0.126893       0.09244
b_priorityLanes       0.162256    0.100970     1.60698    0.097985       1.65592
b_freeParking         0.119343    0.091798     1.30006    0.090073       1.32496
b_tollExemption       0.000000          NA          NA          NA            NA

Number of inter-individual draws            : 1000 (halton)

LL(start)                                   : -4479.28
LL at equal shares, LL(0)                   : -3609.22
LL at observed shares, LL(C)                : -3567.91
LL(final)                                   : -3281.87
Rho-squared vs equal shares                  :  0.0907 
Adj.Rho-squared vs equal shares              :  0.086 
Rho-squared vs observed shares               :  0.0802 
Adj.Rho-squared vs observed shares           :  0.0754 
AIC                                         :  6597.75 
BIC                                         :  6709.23 

#########################################################################################################################
Estimates:
  Estimate        s.e.   t.rat.(0)    Rob.s.e. Rob.t.rat.(0)
mu_log_b_pp          -5.909294    0.645563   -9.153711    0.553016    -10.685577
sigma_log_b_pp        1.562575    0.465600    3.356043    0.320284      4.878713
mu_log_b_oc          -5.800894    0.641728   -9.039488    0.385296    -15.055697
sigma_log_b_oc       -3.488818    0.389908   -8.947788    0.180303    -19.349770
mu_log_b_ct          -7.793861    0.918771   -8.482919    0.523866    -14.877589
sigma_log_b_ct       -8.854007    0.911020   -9.718784    0.390943    -22.647834
b_r                   0.089663    0.027328    3.281034    0.026917      3.331065
b_cf                  0.229010    0.089980    2.545131    0.088143      2.598170
b_e                  -0.162065    0.097198   -1.667369    0.097390     -1.664079
asc_CV                0.000000          NA          NA          NA            NA
asc_EV               -0.205701    0.247575   -0.830863    0.228458     -0.900390
b_Male                0.066888    0.096011    0.696669    0.092058      0.726577
b_female              0.000000          NA          NA          NA            NA
b_ageAvg              0.012195    0.006265    1.946654    0.005925      2.058185
b_dailyDistanceAvg   -0.006898    0.002650   -2.603070    0.002646     -2.607248
b_knowledgeMiddle     0.357275    0.116975    3.054277    0.108323      3.298227
b_knowledgeLow        0.000000          NA          NA          NA            NA
b_knowledgeHigh    -8.3012e-04    0.138163   -0.006008    0.128867     -0.006442
b_priorityLanes       0.165524    0.101480    1.631091    0.098483      1.680735
b_freeParking         0.120359    0.092300    1.304000    0.090571      1.328894
b_tollExemption       0.000000          NA          NA          NA            NA

Number of inter-individual draws            : 1500 (halton)

LL(start)                                   : -4479.28
LL at equal shares, LL(0)                   : -3609.22
LL at observed shares, LL(C)                : -3567.91
LL(final)                                   : -3281.96
Rho-squared vs equal shares                  :  0.0907 
Adj.Rho-squared vs equal shares              :  0.086 
Rho-squared vs observed shares               :  0.0801 
Adj.Rho-squared vs observed shares           :  0.0754 
AIC                                         :  6597.92 
BIC                                         :  6709.4 

#############################################################################################################################
Estimates:
  Estimate        s.e.   t.rat.(0)    Rob.s.e. Rob.t.rat.(0)
mu_log_b_pp          -5.891152         NaN         NaN    0.613070      -9.60926
sigma_log_b_pp        1.568345         NaN         NaN    0.406604       3.85718
mu_log_b_oc          -8.448653         NaN         NaN    0.688475     -12.27155
sigma_log_b_oc       -5.525391         NaN         NaN    0.399658     -13.82530
mu_log_b_ct          -7.504674         NaN         NaN    0.697641     -10.75722
sigma_log_b_ct       -8.434795         NaN         NaN    0.680577     -12.39359
b_r                   0.090022    0.027277     3.30031    0.026866       3.35083
b_cf                  0.233208    0.089695     2.60000    0.087995       2.65024
b_e                  -0.162675    0.096935    -1.67819    0.097226      -1.67316
asc_CV                0.000000          NA          NA          NA            NA
asc_EV               -0.151563    0.240597    -0.62995    0.223916      -0.67687
b_Male                0.068848    0.095285     0.72254    0.091308       0.75402
b_female              0.000000          NA          NA          NA            NA
b_ageAvg              0.012237    0.006206     1.97173    0.005877       2.08206
b_dailyDistanceAvg   -0.006929    0.002614    -2.65080    0.002608      -2.65655
b_knowledgeMiddle     0.358307    0.115738     3.09586    0.106544       3.36298
b_knowledgeLow        0.000000          NA          NA          NA            NA
b_knowledgeHigh       0.004836    0.136609     0.03540    0.127111       0.03804
b_priorityLanes       0.164601    0.101325     1.62449    0.098277       1.67486
b_freeParking         0.117973    0.091999     1.28234    0.090398       1.30505
b_tollExemption       0.000000          NA          NA          NA            NA

Number of inter-individual draws            : 2000 (halton)

LL(start)                                   : -4479.28
LL at equal shares, LL(0)                   : -3609.22
LL at observed shares, LL(C)                : -3567.91
LL(final)                                   : -3281.84
Rho-squared vs equal shares                  :  0.0907 
Adj.Rho-squared vs equal shares              :  0.086 
Rho-squared vs observed shares               :  0.0802 
Adj.Rho-squared vs observed shares           :  0.0754 
AIC                                         :  6597.69 
BIC                                         :  6709.17
Top
stephanehess
Site Admin
Posts: 998
Joined: 24 Apr 2020, 16:29

Re: Result change drastically when increase halton draws from 1500 to 2000 draws

Post by stephanehess »

Hi

more draws is always better in terms of the approximation of the integral. So if the LL changes a lot, then that means that the lower number was too low. Once you use a large enough number, the results will stabilise. Please have a look at http://apollochoicemodelling.com/faq.html and the bit on draws

With too low a number of draws, you might even get "convergence" for a model that is overspecified

Stephane
--------------------------------
Stephane Hess
www.stephanehess.me.uk
Top
Post Reply

Return to “Model estimation”