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.

latent class model with panel term

Ask questions about how to estimate models and how to change your settings for estimation.
Post Reply
lxiaolei
Posts: 7
Joined: 21 Sep 2023, 11:22

latent class model with panel term

Post by lxiaolei »

Hi,

I want to add the panel term in the latent class model, the panel term is assumed to vary over individuals but be constant over choice situation for each individual, the results show that the efficients values of one class are very high while another are normal. Could you please tell me how to add the panel term in the latent class model? Thanks very much!

Code: Select all

### Initialise
rm(list = ls())
library(apollo)

apollo_initialise()

### Set core controls
apollo_control = list(
  modelName  = "social demo as class membership",
  modelDescr = "social demo as class membership",
  indivID    = "ID",
  nCores     = 16,
  outputDirectory = "output"
)
database = read.csv("apollo_data.csv",header=TRUE)
# ################################################################# #
#### DEFINE MODEL PARAMETERS                                     ####
# ################################################################# #

### Vector of parameters, including any that are kept fixed in estimation
apollo_beta = c(# Choice parameters
  asc_1_A               = 0, asc_1_B               = 0, 
  asc_2_A               = 0, asc_2_B               = 0,
  sigma_panel           = 0, 
  b_traveltime_A        = 0, b_traveltime_B        = 0,
  b_sd_A                = 0, b_sd_B                = 0,
  b_crowded_1_A         = 0, b_crowded_1_B         = 0,
  b_crowded_2_A         = 0, b_crowded_2_B         = 0,
  b_crowded_3_A         = 0, b_crowded_3_B         = 0,
  b_cost_A              = 0, b_cost_B              = 0,
  b_transfers_A         = 0, b_transfers_B         = 0,
  b_transfert_A         = 0, b_transfert_B         = 0,
  b_seat_A              = 0, b_seat_B              = 0,
  b_familiar_A          = 0, b_familiar_B          = 0,
  b_tt_pur_A            = 0, b_tt_pur_B            = 0,
  b_tt_pre_A            = 0, b_tt_pre_B            = 0,  
  beta_gender_male_A    = 0, beta_gender_male_B    = 0,
  beta_stay_1_A         = 0, beta_stay_1_B         = 0,
  beta_driver_A         = 0, beta_driver_B         = 0,
  beta_income_1_A       = 0, beta_income_1_B       = 0,
  beta_income_2_A       = 0, beta_income_2_B       = 0,
  beta_age_1_A          = 0, beta_age_1_B          = 0,
  beta_com_metro_A      = 0, beta_com_metro_B       = 0,
  beta_ncom_metro_A     = 0, beta_ncom_metro_B      = 0,
  

  # Class allocation parameters
  delta_A = 0, delta_B = 0)

### Vector with names (in quotes) of parameters to be kept fixed at their starting value in apollo_beta, use apollo_beta_fixed = c() if none
apollo_fixed = c("asc_2_A","asc_2_B","delta_B","beta_gender_male_B","beta_age_1_B","beta_income_1_B","beta_income_2_B","beta_stay_1_B","beta_driver_B","beta_com_metro_B","beta_ncom_metro_B")
# ################################################################# #
#### DEFINE RANDOM COMPONENTS                                    ####
# ################################################################# #
### Set parameters for generating draws
apollo_draws = list(
  interDrawsType = "halton", 
  interNDraws    = 100,
  interNormDraws = c("draws_hsk1","draws_hsk2"))

### Create random parameters
apollo_randCoeff=function(apollo_beta, apollo_inputs){
  randcoeff = list()
  randcoeff[["hsk1"]] = sigma_panel*draws_hsk1
  randcoeff[["hsk2"]] = sigma_panel*draws_hsk2
  
  return(randcoeff)
}
# ################################################################# #
#### DEFINE LATENT CLASS COMPONENTS                              ####
# ################################################################# #

apollo_lcPars=function(apollo_beta, apollo_inputs){
  lcpars = list()
  lcpars[["asc_1"       ]] = list(asc_1_A        , asc_1_B  )
  lcpars[["asc_2"       ]] = list(asc_2_A        , asc_2_B  )
  lcpars[["b_traveltime"]] = list(b_traveltime_A , b_traveltime_B)
  lcpars[["b_sd"        ]] = list(b_sd_A         , b_sd_B        )
  lcpars[["b_crowded_1" ]] = list(b_crowded_1_A  , b_crowded_1_B )
  lcpars[["b_crowded_2" ]] = list(b_crowded_2_A  , b_crowded_2_B )
  lcpars[["b_crowded_3" ]] = list(b_crowded_3_A  , b_crowded_3_B )
  lcpars[["b_cost"      ]] = list(b_cost_A       , b_cost_B      )
  lcpars[["b_transfers" ]] = list(b_transfers_A  , b_transfers_B )
  lcpars[["b_transfert" ]] = list(b_transfert_A  , b_transfert_B )
  lcpars[["b_seat"      ]] = list(b_seat_A       , b_seat_B      )
  lcpars[["b_familiar"  ]] = list(b_familiar_A   , b_familiar_B)  
  lcpars[["b_tt_pur"    ]]  = list(b_tt_pur_A     , b_tt_pur_B)
  lcpars[["b_tt_pre"    ]]  = list(b_tt_pre_A     , b_tt_pre_B)
  
  V = list(class_a = delta_A+beta_gender_male_A*gender_male+beta_age_1_A*age_e+beta_income_1_A*income_1+beta_income_2_A*income_2+beta_stay_1_A*stay_e+beta_driver_A*driver+beta_com_metro_A*com_metro+beta_ncom_metro_A*ncom_metro, 
           class_b = delta_B+beta_gender_male_B*gender_male+beta_age_1_B*age_e+beta_income_1_B*income_1+beta_income_2_B*income_2+beta_stay_1_B*stay_e+beta_driver_B*driver+beta_com_metro_B*com_metro+beta_ncom_metro_B*ncom_metro)
  
  ### Settings for class allocation models
  classAlloc_settings = list(
    classes      = c(class_a=1, class_b=2), 
    utilities    = V  
  )
  
  lcpars[["pi_values"]] = apollo_classAlloc(classAlloc_settings)
  
  return(lcpars)
}
# ################################################################# #
#### GROUP AND VALIDATE INPUTS                                   ####
# ################################################################# #

apollo_inputs = apollo_validateInputs()

# ################################################################# #
#### DEFINE MODEL AND LIKELIHOOD FUNCTION                        ####
# ################################################################# #

apollo_probabilities=function(apollo_beta, apollo_inputs, functionality="estimate"){
  
  ### Initialise
  apollo_attach(apollo_beta, apollo_inputs)
  on.exit(apollo_detach(apollo_beta, apollo_inputs))
  P = list()
  
  mnl_settings = list(
    alternatives = c(alt1=1, alt2=2),
    avail       = list(alt1 = 1, alt2 = 1),
    choiceVar    = best
  )
  
  S <- 2
  for(s in 1:S){
    ### Utilities for alternatives
    V = list()
    V[["alt1"]] = asc_1[[s]] +b_traveltime[[s]]*R1_traveltime+b_sd[[s]]*R1_sd+b_crowded_1[[s]]*R1_crowded_1+ b_crowded_2[[s]]*R1_crowded_2+ b_crowded_3[[s]]*R1_crowded_3+b_cost[[s]]*R1_cost+ b_transfers[[s]]*R1_transfers + b_transfert[[s]]*R1_transtime + b_seat[[s]]*R1_seat + b_familiar[[s]]*R1_familiar+ b_tt_pur[[s]]*R1_traveltime*pur + b_tt_pre[[s]]*R1_traveltime*pre+hsk1
    V[["alt2"]] = asc_2[[s]] +b_traveltime[[s]]*R2_traveltime+b_sd[[s]]*R2_sd+b_crowded_1[[s]]*R2_crowded_1+ b_crowded_2[[s]]*R2_crowded_2+ b_crowded_3[[s]]*R2_crowded_3+b_cost[[s]]*R2_cost+ b_transfers[[s]]*R2_transfers + b_transfert[[s]]*R2_transtime + b_seat[[s]]*R2_seat + b_familiar[[s]]*R2_familiar+ b_tt_pur[[s]]*R2_traveltime*pur + b_tt_pre[[s]]*R2_traveltime*pre+hsk2
    ### Define settings for MNL model component
    mnl_settings$utilities     = V
    mnl_settings$componentName = paste0("Class_",s)
    ### Compute within-class choice probabilities using MNL model
    P[[paste0("Class_",s)]] = apollo_mnl(mnl_settings, functionality)
  } 
  lc_settings  = list(inClassProb = P, classProb=pi_values)
  P[["model"]] = apollo_lc(lc_settings, apollo_inputs, functionality)
  P = apollo_panelProd(P, apollo_inputs, functionality)
  P = apollo_avgInterDraws(P, apollo_inputs, functionality)
  P = apollo_prepareProb(P, apollo_inputs, functionality)
  return(P)
}

# ################################################################# #
#### MODEL ESTIMATION                                            ####
# ################################################################# #
estimate_settings = list(maxIterations = 400)
model = apollo_estimate(apollo_beta, 
                        apollo_fixed, 
                        apollo_probabilities, 
                        apollo_inputs,
                        estimate_settings=estimate_settings)
# ################################################################# #
#### MODEL OUTPUTS                                               ####
# ################################################################# #

# ----------------------------------------------------------------- #
#---- FORMATTED OUTPUT (TO SCREEN)                               ----
# ----------------------------------------------------------------- #

apollo_modelOutput(model)

# ----------------------------------------------------------------- #
#---- FORMATTED OUTPUT (TO FILE, using model name)               ----
# ----------------------------------------------------------------- #

apollo_saveOutput(model)
results:

Code: Select all

                      Estimate        s.e.   t.rat.(0)    Rob.s.e. Rob.t.rat.(0)
asc_1_A             -12.265959  135.887139    -0.09027    1.761668      -6.96270
asc_1_B               0.095975    0.041253     2.32650    0.037517       2.55820
asc_2_A               0.000000          NA          NA          NA            NA
asc_2_B               0.000000          NA          NA          NA            NA
sigma_panel          -0.003908    0.062434    -0.06259    0.002676      -1.46036
b_traveltime_A      -40.069785  417.817473    -0.09590    5.105042      -7.84906
b_traveltime_B       -0.127703    0.008018   -15.92605    0.009797     -13.03485
b_sd_A               -6.150171   59.345161    -0.10363    0.817306      -7.52493
b_sd_B               -0.040716    0.012580    -3.23669    0.012841      -3.17072
b_crowded_1_A       189.510608 1947.191969     0.09733   24.335499       7.78741
b_crowded_1_B         0.552527    0.056367     9.80239    0.061553       8.97642
b_crowded_2_A        40.171986  420.835024     0.09546    4.841213       8.29792
b_crowded_2_B         0.119101    0.060908     1.95545    0.058396       2.03953
b_crowded_3_A       -57.866281  596.305548    -0.09704    7.321333      -7.90379
b_crowded_3_B        -0.347249    0.061037    -5.68920    0.066292      -5.23819
b_cost_A            180.873897 1877.033936     0.09636   23.040848       7.85014
b_cost_B             -0.628583    0.064329    -9.77130    0.062828     -10.00474
b_transfers_A      -157.078402 1625.458541    -0.09664   19.727091      -7.96257
b_transfers_B        -0.402618    0.059174    -6.80399    0.062401      -6.45210
b_transfert_A        10.318179  107.955343     0.09558    1.198330       8.61047
b_transfert_B        -0.065262    0.015727    -4.14971    0.015134      -4.31240
b_seat_A             58.228285  597.252332     0.09749    7.435896       7.83070
b_seat_B              0.487150    0.031054    15.68727    0.035592      13.68722
b_familiar_A         40.000055  418.488401     0.09558    5.096076       7.84919
b_familiar_B          0.164488    0.032391     5.07815    0.035545       4.62756
b_tt_pur_A          -46.937469  488.218721    -0.09614    5.995559      -7.82871
b_tt_pur_B            0.073335    0.011794     6.21784    0.010776       6.80517
b_tt_pre_A           -6.731735   70.533965    -0.09544    0.808283      -8.32844
b_tt_pre_B           -0.061951    0.008757    -7.07414    0.008140      -7.61046
beta_gender_male_A   -0.048532    0.059183    -0.82003    0.062268      -0.77941
beta_gender_male_B    0.000000          NA          NA          NA            NA
beta_stay_1_A         0.123463    0.063634     1.94020    0.072095       1.71251
beta_stay_1_B         0.000000          NA          NA          NA            NA
beta_driver_A        -0.034268    0.073338    -0.46727    0.076106      -0.45027
beta_driver_B         0.000000          NA          NA          NA            NA
beta_income_1_A       0.059468    0.091000     0.65350    0.096040       0.61920
beta_income_1_B       0.000000          NA          NA          NA            NA
beta_income_2_A      -0.007143    0.089019    -0.08024    0.095551      -0.07475
beta_income_2_B       0.000000          NA          NA          NA            NA
beta_age_1_A          0.023061    0.072817     0.31671    0.081670       0.28237
beta_age_1_B          0.000000          NA          NA          NA            NA
beta_com_metro_A     -0.040796    0.073363    -0.55609    0.075496      -0.54038
beta_com_metro_B      0.000000          NA          NA          NA            NA
beta_ncom_metro_A    -0.003066    0.060130    -0.05099    0.064939      -0.04722
beta_ncom_metro_B     0.000000          NA          NA          NA            NA
delta_A              -0.834315    0.129857    -6.42487    0.136619      -6.10690
delta_B               0.000000          NA          NA          NA            NA
Top
stephanehess
Site Admin
Posts: 998
Joined: 24 Apr 2020, 16:29

Re: latent class model with panel term

Post by stephanehess »

Hi

did your model without the panel term work as expected?

Either way, I think you need to do the averaging across draws inside the classes, and also the product across choices (which also should happen inside the class in your model without the error component)

So instead of

Code: Select all

...
    P[[paste0("Class_",s)]] = apollo_mnl(mnl_settings, functionality)
  } 
  lc_settings  = list(inClassProb = P, classProb=pi_values)
  P[["model"]] = apollo_lc(lc_settings, apollo_inputs, functionality)
  P = apollo_panelProd(P, apollo_inputs, functionality)
  P = apollo_avgInterDraws(P, apollo_inputs, functionality)
...
you should have

Code: Select all

...
    P[[paste0("Class_",s)]] = apollo_mnl(mnl_settings, functionality)
    P[[paste0("Class_",s)]] = apollo_panelProd(P[[paste0("Class_",s)]], apollo_inputs, functionality)
    P[[paste0("Class_",s)]] = apollo_avgInterDraws(P[[paste0("Class_",s)]], apollo_inputs, functionality)
  } 
  lc_settings  = list(inClassProb = P, classProb=pi_values)
  P[["model"]] = apollo_lc(lc_settings, apollo_inputs, functionality)
...
--------------------------------
Stephane Hess
www.stephanehess.me.uk
Top
Post Reply

Return to “Model estimation”