Error in ICLV estimation: "Estimation Failed. No covariance matrix to compute."

janak12_jp · Post by **janak12_jp** » 25 Jun 2024, 07:44

I am trying to develop an ICLV model with binary logit. I get the results for parameter estimates but could not compute the covariance matrix and subsequently SE.
Please note that I have tried several specifications with trail & error looking to the large coefficients, changing priors (started with values from simple logit model). For the sake of clarification, I changed the fixed alternative setting as well. But the issue remains the same.

Attaching a simple specified code for reference.

Code: Select all

## Initialization

rm(list=ls())
library(apollo)
apollo_initialise()

## Set Core Controls

apollo_control= list(
  modelName= "ICLV",
  modelDescr= "Mode choice with latent variable",
  indivID= "ID"
 )

## Load database

database= read.csv("#####/iclvdata-2.csv", header=T)

## Parameter operations

database$orgbus= database$orgbus / 100
database$desbus= database$desbus / 100
database$tc_mpv= database$tc_mpv / 1000
database$tc_mpt= database$tc_mpt / 1000
database$tt_pv= database$tt_pv / 10
database$tt_pt= database$tt_pt / 10

## Definition of Model Parameters
### Vector of Parameters to be estimated

apollo_beta= c(# Alternate Specific Constants
               asc_pv= 1.029,
               asc_pt= 0,
                              
               # Social-Economic parameters
               gamma_female_comfort= 0, gamma_female_ttr= 0, gamma_female_cleanliness= 0, gamma_female_staffbehaviour= 0, gamma_female_envisusitain= 0, 
               
               gamma_age_under18_comfort= 0, gamma_age_under18_ttr= 0, gamma_age_under18_cleanliness= 0, gamma_age_under18_staffbehaviour= 0, gamma_age_under18_envisusitain= 0, 
               
               gamma_age_1830_comfort= 0, gamma_age_1830_ttr= 0, gamma_age_1830_cleanliness= 0, gamma_age_1830_staffbehaviour= 0, gamma_age_1830_envisusitain= 0,
               
	       gamma_age_3040_comfort= 0, gamma_age_3040_ttr= 0, gamma_age_3040_cleanliness= 0, gamma_age_3040_staffbehaviour= 0, gamma_age_3040_envisusitain= 0,

	       gamma_age_4050_comfort= 0, gamma_age_4050_ttr= 0, gamma_age_4050_cleanliness= 0, gamma_age_4050_staffbehaviour= 0, gamma_age_4050_envisusitain= 0, 
       
               gamma_age_5060_comfort= 0, gamma_age_5060_ttr= 0, gamma_age_5060_cleanliness= 0, gamma_age_5060_staffbehaviour= 0, gamma_age_5060_envisusitain= 0, 
	      
               gamma_age_above60= 0, #fixed
               gamma_fincome_30k_comfort= 0, gamma_fincome_30k_ttr= 0, gamma_fincome_30k_cleanliness= 0, gamma_fincome_30k_staffbehaviour= 0, gamma_fincome_30k_envisusitain= 0, 
	       
               gamma_fincome_30k60k_comfort= 0, gamma_fincome_30k60k_ttr= 0, gamma_fincome_30k60k_cleanliness= 0, gamma_fincome_30k60k_staffbehaviour= 0, gamma_fincome_30k60k_envisusitain= 0, 

 	       gamma_fincome_60k90k_comfort= 0, gamma_fincome_60k90k_ttr= 0, gamma_fincome_60k90k_cleanliness= 0, gamma_fincome_60k90k_staffbehaviour= 0, gamma_fincome_60k90k_envisusitain= 0, 
               
               gamma_fincome_90k120k_comfort= 0, gamma_fincome_90k120k_ttr= 0, gamma_fincome_90k120k_cleanliness= 0, gamma_fincome_90k120k_staffbehaviour= 0, gamma_fincome_90k120k_envisusitain= 0,
               
               gamma_fincome_120k150k_comfort= 0, gamma_fincome_120k150k_ttr= 0, gamma_fincome_120k150k_cleanliness= 0, gamma_fincome_120k150k_staffbehaviour= 0, gamma_fincome_120k150k_envisusitain= 0,
                                                        
	       gamma_fincome_150k= 0, #fixed
             
               # Trip attributes
               b_tt_pv= -0.403, b_tt_pt= 0, 
               b_tc_pv= 0.008, b_tc_pt= 0, 
               
               # Latent variables and thresholds in SEM component
               zeta_comfort1= 1, zeta_comfort2= 1, zeta_comfort3= 1, zeta_comfort4= 1,
               zeta_ttr1= 1, zeta_ttr2= 1, zeta_ttr3= 1,
               zeta_cleanliness1= 1, zeta_cleanliness2= 1,
	       zeta_staffbehaviour1= 1, zeta_staffbehaviour2= 1,
               zeta_envisusitain1= 1, zeta_envisusitain2= 1,
              
               tau_comfort1_1= -2, tau_comfort1_2= -1, tau_comfort1_3= 1, 
               tau_comfort2_1= -2, tau_comfort2_2= -1, tau_comfort2_3= 1, tau_comfort2_4= 2,
               tau_comfort3_1= -2, tau_comfort3_2= -1, tau_comfort3_3= 1,
               tau_comfort4_1= -2, tau_comfort4_2= -1, tau_comfort4_3= 1,
               tau_ttr1_1= -2, tau_ttr1_2= -1, tau_ttr1_3= 1, 
               tau_ttr2_1= -2, tau_ttr2_2= -1, tau_ttr2_3= 1,
               tau_ttr3_1= -1, tau_ttr3_2= 1,
               tau_cleanliness1_1= -1, tau_cleanliness1_2= 1,
               tau_cleanliness2_1= -2, tau_cleanliness2_2= -1, tau_cleanliness2_3= 1,
               tau_staffbehaviour1_1= -2, tau_staffbehaviour1_2= -1, tau_staffbehaviour1_3= 1, tau_staffbehaviour1_4= 2,
               tau_staffbehaviour2_1= -2, tau_staffbehaviour2_2= -1, tau_staffbehaviour2_3= 1, tau_staffbehaviour2_4= 2,
               tau_envisusitain1_1= -1, tau_envisusitain1_2= 1,
               tau_envisusitain2_1= -2, tau_envisusitain2_2= -1, tau_envisusitain2_3= 1,                            
             
               # Latent variables in choice model component
               lambda_comfort=0, 
               lambda_ttr=0, 
               lambda_cleanliness=0, 
               lambda_staffbehaviour=0, 
               lambda_envisusitain=0, 


               # ETA
               eta1= 0, eta2= 0, eta3= 0, eta4= 0, eta6= 0
               )

## Parameters to be kept FIXED

apollo_fixed= c("asc_pt", 'gamma_age_above60', 'gamma_fincome_150k')


## Group and Validate Inputs

apollo_inputs= apollo_validateInputs()

## Model Definition and Likelihood Function

apollo_probabilities= function(apollo_beta, apollo_inputs, functionality= "estimate"){
  
  # Attach inputs and detach after function exit
  
  apollo_attach(apollo_beta, apollo_inputs)
  on.exit(apollo_detach(apollo_beta, apollo_inputs))
  
  # Create Random LVs
  
	comfort= gamma_female_comfort*female + gamma_age_under18_comfort*(age==1) + gamma_age_1830_comfort*(age==2) + gamma_age_3040_comfort*(age==3) + gamma_age_4050_comfort*(age==4) + gamma_age_5060_comfort*(age==5) + gamma_fincome_30k_comfort*(fincome==1) + gamma_fincome_30k60k_comfort*(fincome==2) + gamma_fincome_60k90k_comfort*(fincome==3) + gamma_fincome_90k120k_comfort*(fincome==4) + gamma_fincome_120k150k_comfort*(fincome==5) + eta1 
       ttr= gamma_female_ttr*female + gamma_age_under18_ttr*(age==1) + gamma_age_1830_ttr*(age==2) + gamma_age_3040_ttr*(age==3) + gamma_age_4050_ttr*(age==4) + gamma_age_5060_ttr*(age==5) + gamma_fincome_30k_ttr*(fincome==1) + gamma_fincome_30k60k_ttr*(fincome==2) + gamma_fincome_60k90k_ttr*(fincome==3) + gamma_fincome_90k120k_ttr*(fincome==4) + gamma_fincome_120k150k_ttr*(fincome==5) + eta2
        cleanliness= gamma_female_cleanliness*female + gamma_age_under18_cleanliness*(age==1) + gamma_age_1830_cleanliness*(age==2) + gamma_age_3040_cleanliness*(age==3) + gamma_age_4050_cleanliness*(age==4) + gamma_age_5060_cleanliness*(age==5) + gamma_fincome_30k_cleanliness*(fincome==1) + gamma_fincome_30k60k_cleanliness*(fincome==2) + gamma_fincome_60k90k_cleanliness*(fincome==3) + gamma_fincome_90k120k_cleanliness*(fincome==4) + gamma_fincome_120k150k_cleanliness*(fincome==5) + eta3
        staffbehaviour= gamma_female_staffbehaviour*female + gamma_age_under18_staffbehaviour*(age==1) + gamma_age_1830_staffbehaviour*(age==2) + gamma_age_3040_staffbehaviour*(age==3) + gamma_age_4050_staffbehaviour*(age==4) + gamma_age_5060_staffbehaviour*(age==5) + gamma_fincome_30k_staffbehaviour*(fincome==1) + gamma_fincome_30k60k_staffbehaviour*(fincome==2) + gamma_fincome_60k90k_staffbehaviour*(fincome==3) + gamma_fincome_90k120k_staffbehaviour*(fincome==4) + gamma_fincome_120k150k_staffbehaviour*(fincome==5) + eta4
       envisusitain= gamma_female_envisusitain*female + gamma_age_under18_envisusitain*(age==1) + gamma_age_1830_envisusitain*(age==2) + gamma_age_3040_envisusitain*(age==3) + gamma_age_4050_envisusitain*(age==4) + gamma_age_5060_envisusitain*(age==5) + gamma_fincome_30k_envisusitain*(fincome==1) + gamma_fincome_30k60k_envisusitain*(fincome==2) + gamma_fincome_60k90k_envisusitain*(fincome==3) + gamma_fincome_90k120k_envisusitain*(fincome==4) + gamma_fincome_120k150k_envisusitain*(fincome==5) + eta6
       
  # Creating list of probabilities P
  
  P= list()
  
  ## Likelihood of Choices
  ## List of utilities 
  
  V= list()

  V[['pv']]= (asc_pv + b_tt_pv*tt_pv + b_tc_pv*tc_mpv +  
                    lambda_comfort*comfort + lambda_ttr*ttr + lambda_cleanliness*cleanliness + lambda_staffbehaviour*staffbehaviour + lambda_envisusitain*envisusitain)
                    
  V[['pt']]= (asc_pt + b_tt_pt*tt_pt + b_tc_pt*tc_mpt + 
                    lambda_comfort*comfort + lambda_ttr*ttr + lambda_cleanliness*cleanliness + lambda_staffbehaviour*staffbehaviour + lambda_envisusitain*envisusitain)
 
    ## Settings for MNL model component
  
  mnl_settings= list(
    alternatives= c(pv=1, pt=2),
    avail= list(pv=av_pv, pt=av_pt),
    choiceVar= mode,
    V= V
  )
  
  ## Computing probabilities for MNL model component
  
  P[["choice"]]= apollo_mnl(mnl_settings, functionality)
  
  # Likelihood of indicators to the LVs
  
  ol_settings1= list(outcomeOrdered= comfort1,
                     V= zeta_comfort1*comfort,
                     tau= list(tau_comfort1_1, tau_comfort1_2, tau_comfort1_3),
                     rows=comfort1!=0)
  ol_settings2= list(outcomeOrdered= comfort2,
                     V= zeta_comfort2*comfort,
                     tau= list(tau_comfort2_1, tau_comfort2_2, tau_comfort2_3, tau_comfort2_4),
                     rows=comfort2!=0)
  ol_settings3= list(outcomeOrdered= comfort3,
                     V= zeta_comfort3*comfort,
                     tau= list(tau_comfort3_1, tau_comfort3_2, tau_comfort3_3),
                     rows=comfort3!=0)
  ol_settings4= list(outcomeOrdered= comfort4,
                     V= zeta_comfort4*comfort,
                     tau= list(tau_comfort4_1, tau_comfort4_2, tau_comfort4_3),
                     rows=comfort4!=0)
  ol_settings5= list(outcomeOrdered= ttr1,
                     V= zeta_ttr1*ttr,
                     tau= list(tau_ttr1_1, tau_ttr1_2, tau_ttr1_3),
                     rows=ttr1!=0)
  ol_settings6= list(outcomeOrdered= ttr2,
                     V= zeta_ttr2*ttr,
                     tau= list(tau_ttr2_1, tau_ttr2_2, tau_ttr2_3),
                     rows=ttr2!=0)
  ol_settings7= list(outcomeOrdered= ttr3,
                     V= zeta_ttr3*ttr,
                     tau= list(tau_ttr3_1, tau_ttr3_2),
                     rows=ttr3!=0)
  ol_settings8= list(outcomeOrdered= cleanliness1,
                     V= zeta_cleanliness1*cleanliness,
                     tau= list(tau_cleanliness1_1, tau_cleanliness1_2),
                     rows=cleanliness1!=0)
  ol_settings9= list(outcomeOrdered= cleanliness2,
                     V= zeta_cleanliness2*cleanliness,
                     tau= list(tau_cleanliness2_1, tau_cleanliness2_2, tau_cleanliness2_3),
                     rows=cleanliness2!=0)
  ol_settings10= list(outcomeOrdered= staffbehaviour1,
                     V= zeta_staffbehaviour1*staffbehaviour,
                     tau= list(tau_staffbehaviour1_1, tau_staffbehaviour1_2, tau_staffbehaviour1_3, tau_staffbehaviour1_4),
                     rows=staffbehaviour1!=0)
  ol_settings11= list(outcomeOrdered= staffbehaviour2,
                     V= zeta_staffbehaviour2*staffbehaviour,
                     tau= list(tau_staffbehaviour2_1, tau_staffbehaviour2_2, tau_staffbehaviour2_3, tau_staffbehaviour2_4),
                     rows=staffbehaviour2!=0)
   ol_settings14= list(outcomeOrdered= envisusitain1,
                     V= zeta_envisusitain1*envisusitain,
                     tau= list(tau_envisusitain1_1, tau_envisusitain1_2),
                     rows=envisusitain1!=0)
  ol_settings15= list(outcomeOrdered= envisusitain2,
                     V= zeta_envisusitain2*envisusitain,
                     tau= list(tau_envisusitain2_1, tau_envisusitain2_2, tau_envisusitain2_3),
                     rows=envisusitain2!=0)
 
  
  P[["indic_comfort1"]]= apollo_ol(ol_settings1, functionality)
  P[["indic_comfort2"]]= apollo_ol(ol_settings2, functionality)
  P[["indic_comfort3"]]= apollo_ol(ol_settings3, functionality)
  P[["indic_comfort4"]]= apollo_ol(ol_settings4, functionality)
  P[["indic_ttr1"]]= apollo_ol(ol_settings5, functionality)
  P[["indic_ttr2"]]= apollo_ol(ol_settings6, functionality)
  P[["indic_ttr3"]]= apollo_ol(ol_settings7, functionality)
  P[["indic_cleanliness1"]]= apollo_ol(ol_settings8, functionality)
  P[["indic_cleanliness2"]]= apollo_ol(ol_settings9, functionality)
  P[["indic_staffbehaviour1"]]= apollo_ol(ol_settings10, functionality)
  P[["indic_staffbehaviour2"]]= apollo_ol(ol_settings11, functionality)
  P[["indic_envisusitain1"]]= apollo_ol(ol_settings14, functionality)
  P[["indic_envisusitain2"]]= apollo_ol(ol_settings15, functionality)
 
   
  ## Likelihood of the whole combined model
  
  P= apollo_combineModels(P, apollo_inputs, functionality)
    
  ## Prepare and return the outputs of the function
  
  P= apollo_prepareProb(P, apollo_inputs, functionality)
  
  return(P)
}

L <- apollo_probabilities(apollo_beta, apollo_inputs, 'estimate')									
## Model Estimation

# Likelihood before model estimation

apollo_llCalc(apollo_beta, apollo_probabilities, apollo_inputs)

# Estimate Model

model = apollo_estimate(apollo_beta, apollo_fixed, apollo_probabilities, apollo_inputs)

## Model Output

## apollo_modelOutput(model)

apollo_saveOutput(model)

Post by **stephanehess** » 03 Jul 2024, 23:25

Hi

we cannot really help without seeing results/outputs

But your model does not seem to be an ICLV model anyway as you're not using any random components. Did you look at the examples online?

Stephane

janak12_jp · Post by **janak12_jp** » 05 Jul 2024, 06:12

Hi Stephen,

Thank you for the reply. Sorry for posting different code. Please review this code.
I could handle previous one. But the error seems to be different now.

Error in apollo_checkArguments(apollo_probabilities, apollo_randCoeff, :
SYNTAX ISSUE - The arguments for apollo_randCoeff need to be apollo_beta and apollo_inputs
In addition: Warning message:
In names(arguments) == c("apollo_beta", "apollo_inputs") :
longer object length is not a multiple of shorter object length

Code: Select all

## Initialization

rm(list=ls())
library(apollo)
apollo_initialise()

## Set Core Controls

apollo_control= list(
  modelName= "ICLV",
  modelDescr= "Mode choice with latent variable",
  indivID= "ID"
)

## Load database

database= read.csv("~iclvdata-2.csv", header=T)

## Parameter operations

database$orgbus= database$orgbus / 100
database$desbus= database$desbus / 100
database$tc_mpv= database$tc_mpv / 1000
database$tc_mpt= database$tc_mpt / 1000
database$tt_pv= database$tt_pv / 10
database$tt_pt= database$tt_pt / 10

## Definition of Model Parameters
### Vector of Parameters to be estimated

apollo_beta= c(# Alternate Specific Constants
  asc_pv= -1.029,
  asc_pt= 0,
  
  
  # Social-Economic parameters
  gamma_female_comfort= 0, gamma_female_ttr= 0, gamma_female_cleanliness= 0, gamma_female_staffbehaviour= 0, gamma_female_envisusitain= 0, 
  
  gamma_age_under18_comfort= 0, gamma_age_under18_ttr= 0, gamma_age_under18_cleanliness= 0, gamma_age_under18_staffbehaviour= 0, gamma_age_under18_envisusitain= 0, 
  
  gamma_age_1830_comfort= 0, gamma_age_1830_ttr= 0, gamma_age_1830_cleanliness= 0, gamma_age_1830_staffbehaviour= 0, gamma_age_1830_envisusitain= 0,
  
  gamma_age_3040_comfort= 0, gamma_age_3040_ttr= 0, gamma_age_3040_cleanliness= 0, gamma_age_3040_staffbehaviour= 0, gamma_age_3040_envisusitain= 0,
  
  gamma_age_4050_comfort= 0, gamma_age_4050_ttr= 0, gamma_age_4050_cleanliness= 0, gamma_age_4050_staffbehaviour= 0, gamma_age_4050_envisusitain= 0, 
  
  gamma_age_5060_comfort= 0, gamma_age_5060_ttr= 0, gamma_age_5060_cleanliness= 0, gamma_age_5060_staffbehaviour= 0, gamma_age_5060_envisusitain= 0, 
  
  gamma_age_above60= 0, #fixed
  
  gamma_fincome_30k_comfort= 0, gamma_fincome_30k_ttr= 0, gamma_fincome_30k_cleanliness= 0, gamma_fincome_30k_staffbehaviour= 0, gamma_fincome_30k_envisusitain= 0, 
  
  gamma_fincome_30k60k_comfort= 0, gamma_fincome_30k60k_ttr= 0, gamma_fincome_30k60k_cleanliness= 0, gamma_fincome_30k60k_staffbehaviour= 0, gamma_fincome_30k60k_envisusitain= 0, 
  
  gamma_fincome_60k90k_comfort= 0, gamma_fincome_60k90k_ttr= 0, gamma_fincome_60k90k_cleanliness= 0, gamma_fincome_60k90k_staffbehaviour= 0, gamma_fincome_60k90k_envisusitain= 0, 
  
  gamma_fincome_90k120k_comfort= 0, gamma_fincome_90k120k_ttr= 0, gamma_fincome_90k120k_cleanliness= 0, gamma_fincome_90k120k_staffbehaviour= 0, gamma_fincome_90k120k_envisusitain= 0,
  
  gamma_fincome_120k150k_comfort= 0, gamma_fincome_120k150k_ttr= 0, gamma_fincome_120k150k_cleanliness= 0, gamma_fincome_120k150k_staffbehaviour= 0, gamma_fincome_120k150k_envisusitain= 0,
  
  gamma_fincome_150k= 0, #fixed
  
  # Trip attributes
  b_tt_pv= -0.403, b_tt_pt= 0, 
  b_tc_pv= 0.008, b_tc_pt= 0,
  
  # Latent variables and thresholds in SEM component
  zeta_comfort1= 1, zeta_comfort2= 1, zeta_comfort3= 1, zeta_comfort4= 1,
  zeta_ttr1= 1, zeta_ttr2= 1, zeta_ttr3= 1,
  zeta_cleanliness1= 1, zeta_cleanliness2= 1,
  zeta_staffbehaviour1= 1, zeta_staffbehaviour2= 1,
  zeta_envisusitain1= 1, zeta_envisusitain2= 1,
  
  tau_comfort1_1= -2, tau_comfort1_2= -1, tau_comfort1_3= 1, 
  tau_comfort2_1= -2, tau_comfort2_2= -1, tau_comfort2_3= 1, tau_comfort2_4= 2,
  tau_comfort3_1= -2, tau_comfort3_2= -1, tau_comfort3_3= 1,
  tau_comfort4_1= -2, tau_comfort4_2= -1, tau_comfort4_3= 1,
  tau_ttr1_1= -2, tau_ttr1_2= -1, tau_ttr1_3= 1, 
  tau_ttr2_1= -2, tau_ttr2_2= -1, tau_ttr2_3= 1,
  tau_ttr3_1= -1, tau_ttr3_2= 1,
  tau_cleanliness1_1= -1, tau_cleanliness1_2= 1,
  tau_cleanliness2_1= -2, tau_cleanliness2_2= -1, tau_cleanliness2_3= 1,
  tau_staffbehaviour1_1= -2, tau_staffbehaviour1_2= -1, tau_staffbehaviour1_3= 1, tau_staffbehaviour1_4= 2,
  tau_staffbehaviour2_1= -2, tau_staffbehaviour2_2= -1, tau_staffbehaviour2_3= 1, tau_staffbehaviour2_4= 2,
  tau_envisusitain1_1= -1, tau_envisusitain1_2= 1,
  tau_envisusitain2_1= -2, tau_envisusitain2_2= -1, tau_envisusitain2_3= 1,                            
  
  # Latent variables in choice model component
  lambda_comfort=1, 
  lambda_ttr=1, 
  lambda_cleanliness=1, 
  lambda_staffbehaviour=1, 
  lambda_envisusitain=1 
  
)

## Parameters to be kept FIXED

apollo_fixed= c("asc_pt", 'gamma_age_above60', 'gamma_fincome_150k')
apollo_draws = list(
  interDrawsType="halton", 
  interNDraws=500,          
  interUnifDraws=c(),      
  interNormDraws=c("eta1", "eta2", "eta3", "eta4", "eta5"), 
  
  intraDrawsType="",
  intraNDraws=0,          
  intraUnifDraws=c(),     
  intraNormDraws=c()      
)
# Create Random LVs

apollo_randCoeff=function(apollo_beta, apollo_inputs, functionality= "estimate"){
  randcoeff = list()  
  
  randcoeff[["comfort"]]= gamma_female_comfort*female + gamma_age_under18_comfort*(age==1) + gamma_age_1830_comfort*(age==2) + gamma_age_3040_comfort*(age==3) + gamma_age_4050_comfort*(age==4) + gamma_age_5060_comfort*(age==5) + gamma_fincome_30k_comfort*(fincome==1) + gamma_fincome_30k60k_comfort*(fincome==2) + gamma_fincome_60k90k_comfort*(fincome==3) + gamma_fincome_90k120k_comfort*(fincome==4) + gamma_fincome_120k150k_comfort*(fincome==5) + eta1 
  randcoeff[["ttr"]]= gamma_female_ttr*female + gamma_age_under18_ttr*(age==1) + gamma_age_1830_ttr*(age==2) + gamma_age_3040_ttr*(age==3) + gamma_age_4050_ttr*(age==4) + gamma_age_5060_ttr*(age==5) + gamma_fincome_30k_ttr*(fincome==1) + gamma_fincome_30k60k_ttr*(fincome==2) + gamma_fincome_60k90k_ttr*(fincome==3) + gamma_fincome_90k120k_ttr*(fincome==4) + gamma_fincome_120k150k_ttr*(fincome==5) + eta2
  randcoeff[["cleanliness"]]= gamma_female_cleanliness*female + gamma_age_under18_cleanliness*(age==1) + gamma_age_1830_cleanliness*(age==2) + gamma_age_3040_cleanliness*(age==3) + gamma_age_4050_cleanliness*(age==4) + gamma_age_5060_cleanliness*(age==5) + gamma_fincome_30k_cleanliness*(fincome==1) + gamma_fincome_30k60k_cleanliness*(fincome==2) + gamma_fincome_60k90k_cleanliness*(fincome==3) + gamma_fincome_90k120k_cleanliness*(fincome==4) + gamma_fincome_120k150k_cleanliness*(fincome==5) + eta3
  randcoeff[["staffbehaviour"]]= gamma_female_staffbehaviour*female + gamma_age_under18_staffbehaviour*(age==1) + gamma_age_1830_staffbehaviour*(age==2) + gamma_age_3040_staffbehaviour*(age==3) + gamma_age_4050_staffbehaviour*(age==4) + gamma_age_5060_staffbehaviour*(age==5) + gamma_fincome_30k_staffbehaviour*(fincome==1) + gamma_fincome_30k60k_staffbehaviour*(fincome==2) + gamma_fincome_60k90k_staffbehaviour*(fincome==3) + gamma_fincome_90k120k_staffbehaviour*(fincome==4) + gamma_fincome_120k150k_staffbehaviour*(fincome==5) + eta4
  randcoeff[["envisusitain"]]= gamma_female_envisusitain*female + gamma_age_under18_envisusitain*(age==1) + gamma_age_1830_envisusitain*(age==2) + gamma_age_3040_envisusitain*(age==3) + gamma_age_4050_envisusitain*(age==4) + gamma_age_5060_envisusitain*(age==5) + gamma_fincome_30k_envisusitain*(fincome==1) + gamma_fincome_30k60k_envisusitain*(fincome==2) + gamma_fincome_60k90k_envisusitain*(fincome==3) + gamma_fincome_90k120k_envisusitain*(fincome==4) + gamma_fincome_120k150k_envisusitain*(fincome==5) + eta5
  
  
  return(randcoeff)
}

## Group and Validate Inputs

apollo_inputs= apollo_validateInputs()

## Model Definition and Likelihood Function

apollo_probabilities= function(apollo_beta, apollo_inputs, functionality= "estimate"){
  
  # Attach inputs and detach after function exit
  
  apollo_attach(apollo_beta, apollo_inputs)
  on.exit(apollo_detach(apollo_beta, apollo_inputs))
  
  
  # Creating list of probabilities P
  
  P = list()
  
  # Likelihood of indicators to the LVs
  
  ol_settings1= list(outcomeOrdered= comfort1,
                     V= zeta_comfort1*comfort,
                     tau= list(tau_comfort1_1, tau_comfort1_2, tau_comfort1_3),
                     rows=comfort1!=0)
  ol_settings2= list(outcomeOrdered= comfort2,
                     V= zeta_comfort2*comfort,
                     tau= list(tau_comfort2_1, tau_comfort2_2, tau_comfort2_3, tau_comfort2_4),
                     rows=comfort2!=0)
  ol_settings3= list(outcomeOrdered= comfort3,
                     V= zeta_comfort3*comfort,
                     tau= list(tau_comfort3_1, tau_comfort3_2, tau_comfort3_3),
                     rows=comfort3!=0)
  ol_settings4= list(outcomeOrdered= comfort4,
                     V= zeta_comfort4*comfort,
                     tau= list(tau_comfort4_1, tau_comfort4_2, tau_comfort4_3),
                     rows=comfort4!=0)
  ol_settings5= list(outcomeOrdered= ttr1,
                     V= zeta_ttr1*ttr,
                     tau= list(tau_ttr1_1, tau_ttr1_2, tau_ttr1_3),
                     rows=ttr1!=0)
  ol_settings6= list(outcomeOrdered= ttr2,
                     V= zeta_ttr2*ttr,
                     tau= list(tau_ttr2_1, tau_ttr2_2, tau_ttr2_3),
                     rows=ttr2!=0)
  ol_settings7= list(outcomeOrdered= ttr3,
                     V= zeta_ttr3*ttr,
                     tau= list(tau_ttr3_1, tau_ttr3_2),
                     rows=ttr3!=0)
  ol_settings8= list(outcomeOrdered= cleanliness1,
                     V= zeta_cleanliness1*cleanliness,
                     tau= list(tau_cleanliness1_1, tau_cleanliness1_2),
                     rows=cleanliness1!=0)
  ol_settings9= list(outcomeOrdered= cleanliness2,
                     V= zeta_cleanliness2*cleanliness,
                     tau= list(tau_cleanliness2_1, tau_cleanliness2_2, tau_cleanliness2_3),
                     rows=cleanliness2!=0)
  ol_settings10= list(outcomeOrdered= staffbehaviour1,
                      V= zeta_staffbehaviour1*staffbehaviour,
                      tau= list(tau_staffbehaviour1_1, tau_staffbehaviour1_2, tau_staffbehaviour1_3, tau_staffbehaviour1_4),
                      rows=staffbehaviour1!=0)
  ol_settings11= list(outcomeOrdered= staffbehaviour2,
                      V= zeta_staffbehaviour2*staffbehaviour,
                      tau= list(tau_staffbehaviour2_1, tau_staffbehaviour2_2, tau_staffbehaviour2_3, tau_staffbehaviour2_4),
                      rows=staffbehaviour2!=0)
  ol_settings14= list(outcomeOrdered= envisusitain1,
                      V= zeta_envisusitain1*envisusitain,
                      tau= list(tau_envisusitain1_1, tau_envisusitain1_2),
                      rows=envisusitain1!=0)
  ol_settings15= list(outcomeOrdered= envisusitain2,
                      V= zeta_envisusitain2*envisusitain,
                      tau= list(tau_envisusitain2_1, tau_envisusitain2_2, tau_envisusitain2_3),
                      rows=envisusitain2!=0)
  
  
  P[["indic_comfort1"]]= apollo_ol(ol_settings1, functionality)
  P[["indic_comfort2"]]= apollo_ol(ol_settings2, functionality)
  P[["indic_comfort3"]]= apollo_ol(ol_settings3, functionality)
  P[["indic_comfort4"]]= apollo_ol(ol_settings4, functionality)
  P[["indic_ttr1"]]= apollo_ol(ol_settings5, functionality)
  P[["indic_ttr2"]]= apollo_ol(ol_settings6, functionality)
  P[["indic_ttr3"]]= apollo_ol(ol_settings7, functionality)
  P[["indic_cleanliness1"]]= apollo_ol(ol_settings8, functionality)
  P[["indic_cleanliness2"]]= apollo_ol(ol_settings9, functionality)
  P[["indic_staffbehaviour1"]]= apollo_ol(ol_settings10, functionality)
  P[["indic_staffbehaviour2"]]= apollo_ol(ol_settings11, functionality)
  P[["indic_envisusitain1"]]= apollo_ol(ol_settings14, functionality)
  P[["indic_envisusitain2"]]= apollo_ol(ol_settings15, functionality)
  
  ## List of utilities 
  
  V= list()
  
  V[['pv']]= (asc_pv + b_tt_pv*tt_pv + b_tc_pv*tc_mpv +  
                lambda_comfort*comfort + lambda_ttr*ttr + lambda_cleanliness*cleanliness + lambda_staffbehaviour*staffbehaviour + lambda_envisusitain*envisusitain)
  
  V[['pt']]= (asc_pt + b_tt_pt*tt_pt + b_tc_pt*tc_mpt + 
                lambda_comfort*comfort + lambda_ttr*ttr + lambda_cleanliness*cleanliness + lambda_staffbehaviour*staffbehaviour + lambda_envisusitain*envisusitain)
  
  ## Settings for MNL model component
  
  mnl_settings= list(
    alternatives= c(pv=1, pt=2),
    avail= list(pv=av_pv, pt=av_pt),
    choiceVar= mode,
    V= V
  )
  ### Compute probabilities for MNL model component
  
  P[["choice"]] = apollo_mnl(mnl_settings, functionality)
  
  ## Likelihood of the whole combined model
  
  P= apollo_combineModels(P, apollo_inputs, functionality)
  
  ### Average across inter-individual draws
  P = apollo_avgInterDraws(P, apollo_inputs, functionality)
  
  ## Prepare and return the outputs of the function
  
  P= apollo_prepareProb(P, apollo_inputs, functionality)
  
  return(P)
}

L <- apollo_probabilities(apollo_beta, apollo_inputs, 'estimate')									
## Model Estimation

# Likelihood before model estimation

apollo_llCalc(apollo_beta, apollo_probabilities, apollo_inputs)

# Estimate Model

model = apollo_estimate(apollo_beta, apollo_fixed, apollo_probabilities, apollo_inputs)

## Model Output

## apollo_modelOutput(model)

apollo_saveOutput(model)

Post by **stephanehess** » 08 Jul 2024, 07:20

Hi

please have a look at the example files and the manual. You have changed the syntax.

You have

Code: Select all

apollo_randCoeff=function(apollo_beta, apollo_inputs, functionality= "estimate")

instead of

Code: Select all

apollo_randCoeff = function(apollo_beta, apollo_inputs)

Stephane

janak12_jp · Post by **janak12_jp** » 09 Jul 2024, 16:19

Hi, Stephane
Thank you for addressing the issue.
However, the error still persist:

Error in apollo_estimate(apollo_beta, apollo_fixed, apollo_probabilities, :
SPECIFICATION ISSUE - Parameter lambda_comfort does not influence the log-likelihood of your model!

When I remove related specification, the error jumps to next latent variable.

Here is the code:

Code: Select all


## Initialization

rm(list=ls())
library(apollo)
apollo_initialise()

## Set Core Controls

apollo_control= list(
  modelName= "ICLV",
  modelDescr= "Mode choice with latent variable",
  indivID= "ID"
)

## Load database

database= read.csv("C:/Users/janak/Desktop/iclvdata-2.csv", header=T)

## Parameter operations

database$orgbus= database$orgbus / 100
database$desbus= database$desbus / 100
database$tc_mpv= database$tc_mpv / 1000
database$tc_mpt= database$tc_mpt / 1000
database$tt_pv= database$tt_pv / 10
database$tt_pt= database$tt_pt / 10

## Definition of Model Parameters
### Vector of Parameters to be estimated

apollo_beta= c(# Alternate Specific Constants
  asc_pv= 0,
  asc_pt= -1,
  
  
  # Social-Economic parameters
  gamma_female_comfort= 0, gamma_female_ttr= 0, gamma_female_cleanliness= 0, gamma_female_staffbehaviour= 0, gamma_female_envisusitain= 0, 
  
  gamma_age_under18_comfort= 0, gamma_age_under18_ttr= 0, gamma_age_under18_cleanliness= 0, gamma_age_under18_staffbehaviour= 0, gamma_age_under18_envisusitain= 0, 
  
  gamma_age_1830_comfort= 0, gamma_age_1830_ttr= 0, gamma_age_1830_cleanliness= 0, gamma_age_1830_staffbehaviour= 0, gamma_age_1830_envisusitain= 0,
  
  gamma_age_3040_comfort= 0, gamma_age_3040_ttr= 0, gamma_age_3040_cleanliness= 0, gamma_age_3040_staffbehaviour= 0, gamma_age_3040_envisusitain= 0,
  
  gamma_age_4050_comfort= 0, gamma_age_4050_ttr= 0, gamma_age_4050_cleanliness= 0, gamma_age_4050_staffbehaviour= 0, gamma_age_4050_envisusitain= 0, 
  
  gamma_age_5060_comfort= 0, gamma_age_5060_ttr= 0, gamma_age_5060_cleanliness= 0, gamma_age_5060_staffbehaviour= 0, gamma_age_5060_envisusitain= 0, 
  
  gamma_age_above60= 0, #fixed
  
  gamma_fincome_30k_comfort= 0, gamma_fincome_30k_ttr= 0, gamma_fincome_30k_cleanliness= 0, gamma_fincome_30k_staffbehaviour= 0, gamma_fincome_30k_envisusitain= 0, 
  
  gamma_fincome_30k60k_comfort= 0, gamma_fincome_30k60k_ttr= 0, gamma_fincome_30k60k_cleanliness= 0, gamma_fincome_30k60k_staffbehaviour= 0, gamma_fincome_30k60k_envisusitain= 0, 
  
  gamma_fincome_60k90k_comfort= 0, gamma_fincome_60k90k_ttr= 0, gamma_fincome_60k90k_cleanliness= 0, gamma_fincome_60k90k_staffbehaviour= 0, gamma_fincome_60k90k_envisusitain= 0, 
  
  gamma_fincome_90k120k_comfort= 0, gamma_fincome_90k120k_ttr= 0, gamma_fincome_90k120k_cleanliness= 0, gamma_fincome_90k120k_staffbehaviour= 0, gamma_fincome_90k120k_envisusitain= 0,
  
  gamma_fincome_120k150k_comfort= 0, gamma_fincome_120k150k_ttr= 0, gamma_fincome_120k150k_cleanliness= 0, gamma_fincome_120k150k_staffbehaviour= 0, gamma_fincome_120k150k_envisusitain= 0,
  
  gamma_fincome_150k= 0, #fixed
  
  # Trip attributes
  b_tt_pv= -0.403, b_tt_pt= 0, 
  b_tc_pv= 0.008, b_tc_pt= 0,
  
  # Latent variables and thresholds in SEM component
  zeta_comfort1= 1, zeta_comfort2= 1, zeta_comfort3= 1, zeta_comfort4= 1,
  zeta_ttr1= 1, zeta_ttr2= 1, zeta_ttr3= 1,
  zeta_cleanliness1= 1, zeta_cleanliness2= 1,
  zeta_staffbehaviour1= 1, zeta_staffbehaviour2= 1,
  zeta_envisusitain1= 1, zeta_envisusitain2= 1,
  
  tau_comfort1_1= -2, tau_comfort1_2= -1, tau_comfort1_3= 1, 
  tau_comfort2_1= -2, tau_comfort2_2= -1, tau_comfort2_3= 1, tau_comfort2_4= 2,
  tau_comfort3_1= -2, tau_comfort3_2= -1, tau_comfort3_3= 1,
  tau_comfort4_1= -2, tau_comfort4_2= -1, tau_comfort4_3= 1,
  tau_ttr1_1= -2, tau_ttr1_2= -1, tau_ttr1_3= 1, 
  tau_ttr2_1= -2, tau_ttr2_2= -1, tau_ttr2_3= 1,
  tau_ttr3_1= -1, tau_ttr3_2= 1,
  tau_cleanliness1_1= -1, tau_cleanliness1_2= 1,
  tau_cleanliness2_1= -2, tau_cleanliness2_2= -1, tau_cleanliness2_3= 1,
  tau_staffbehaviour1_1= -2, tau_staffbehaviour1_2= -1, tau_staffbehaviour1_3= 1, tau_staffbehaviour1_4= 2,
  tau_staffbehaviour2_1= -2, tau_staffbehaviour2_2= -1, tau_staffbehaviour2_3= 1, tau_staffbehaviour2_4= 2,
  tau_envisusitain1_1= -1, tau_envisusitain1_2= 1,
  tau_envisusitain2_1= -2, tau_envisusitain2_2= -1, tau_envisusitain2_3= 1,                            
  
  # Latent variables in choice model component
  lambda_comfort=0.1, 
  lambda_ttr=0.1, 
  lambda_cleanliness=0.1, 
  lambda_staffbehaviour=0.1, 
  lambda_envisusitain=0.1 
)

## Parameters to be kept FIXED

apollo_fixed= c("asc_pv", 'gamma_age_above60', 'gamma_fincome_150k')

### Set parameters for generating draws
apollo_draws = list(
  interDrawsType="halton", 
  interNDraws=500,          
  interUnifDraws=c(),      
  interNormDraws=c("eta1", "eta2", "eta3", "eta4", "eta5"), 
  
  intraDrawsType="",
  intraNDraws=0,          
  intraUnifDraws=c(),     
  intraNormDraws=c()      
)

# Create Random LVs

apollo_randCoeff=function(apollo_beta, apollo_inputs){
  randcoeff = list()  
  
  randcoeff[["comfort"]]= (gamma_female_comfort*female + gamma_age_under18_comfort*(age==1) + gamma_age_1830_comfort*(age==2) + gamma_age_3040_comfort*(age==3) + gamma_age_4050_comfort*(age==4) + gamma_age_5060_comfort*(age==5) + gamma_fincome_30k_comfort*(fincome==1) + gamma_fincome_30k60k_comfort*(fincome==2) + gamma_fincome_60k90k_comfort*(fincome==3) + gamma_fincome_90k120k_comfort*(fincome==4) + gamma_fincome_120k150k_comfort*(fincome==5) + eta1) 
  randcoeff[["ttr"]]= (gamma_female_ttr*female + gamma_age_under18_ttr*(age==1) + gamma_age_1830_ttr*(age==2) + gamma_age_3040_ttr*(age==3) + gamma_age_4050_ttr*(age==4) + gamma_age_5060_ttr*(age==5) + gamma_fincome_30k_ttr*(fincome==1) + gamma_fincome_30k60k_ttr*(fincome==2) + gamma_fincome_60k90k_ttr*(fincome==3) + gamma_fincome_90k120k_ttr*(fincome==4) + gamma_fincome_120k150k_ttr*(fincome==5) + eta2)
  randcoeff[["cleanliness"]]= (gamma_female_cleanliness*female + gamma_age_under18_cleanliness*(age==1) + gamma_age_1830_cleanliness*(age==2) + gamma_age_3040_cleanliness*(age==3) + gamma_age_4050_cleanliness*(age==4) + gamma_age_5060_cleanliness*(age==5) + gamma_fincome_30k_cleanliness*(fincome==1) + gamma_fincome_30k60k_cleanliness*(fincome==2) + gamma_fincome_60k90k_cleanliness*(fincome==3) + gamma_fincome_90k120k_cleanliness*(fincome==4) + gamma_fincome_120k150k_cleanliness*(fincome==5) + eta3)
  randcoeff[["staffbehaviour"]]= (gamma_female_staffbehaviour*female + gamma_age_under18_staffbehaviour*(age==1) + gamma_age_1830_staffbehaviour*(age==2) + gamma_age_3040_staffbehaviour*(age==3) + gamma_age_4050_staffbehaviour*(age==4) + gamma_age_5060_staffbehaviour*(age==5) + gamma_fincome_30k_staffbehaviour*(fincome==1) + gamma_fincome_30k60k_staffbehaviour*(fincome==2) + gamma_fincome_60k90k_staffbehaviour*(fincome==3) + gamma_fincome_90k120k_staffbehaviour*(fincome==4) + gamma_fincome_120k150k_staffbehaviour*(fincome==5) + eta4)
  randcoeff[["envisusitain"]]= (gamma_female_envisusitain*female + gamma_age_under18_envisusitain*(age==1) + gamma_age_1830_envisusitain*(age==2) + gamma_age_3040_envisusitain*(age==3) + gamma_age_4050_envisusitain*(age==4) + gamma_age_5060_envisusitain*(age==5) + gamma_fincome_30k_envisusitain*(fincome==1) + gamma_fincome_30k60k_envisusitain*(fincome==2) + gamma_fincome_60k90k_envisusitain*(fincome==3) + gamma_fincome_90k120k_envisusitain*(fincome==4) + gamma_fincome_120k150k_envisusitain*(fincome==5) + eta5)
  
  
  return(randcoeff)
}


## Group and Validate Inputs

apollo_inputs= apollo_validateInputs()

## Model Definition and Likelihood Function

apollo_probabilities= function(apollo_beta, apollo_inputs, functionality= "estimate"){
  
  # Attach inputs and detach after function exit
  
  apollo_attach(apollo_beta, apollo_inputs)
  on.exit(apollo_detach(apollo_beta, apollo_inputs))
  
  # Creating list of probabilities P
  
  P= list()
  
  ## Likelihood of Choices
  ## List of utilities 
  
  V= list()
  
  V[['pv']]= (asc_pv + b_tt_pv*tt_pv + b_tc_pv*tc_mpv +  
                lambda_comfort*comfort + lambda_ttr*ttr + lambda_cleanliness*cleanliness + lambda_staffbehaviour*staffbehaviour + lambda_envisusitain*envisusitain)
  
  V[['pt']]= (asc_pt + b_tt_pt*tt_pt + b_tc_pt*tc_mpt + 
                lambda_comfort*comfort + lambda_ttr*ttr + lambda_cleanliness*cleanliness + lambda_staffbehaviour*staffbehaviour + lambda_envisusitain*envisusitain)
  
  ## Settings for MNL model component
  
  mnl_settings= list(
    alternatives= c(pv=1, pt=2),
    avail= list(pv=av_pv, pt=av_pt),
    choiceVar= mode,
    utilities = V,
    componentName = "choice"
  )
  
  
  # Likelihood of indicators to the LVs
  
  ol_settings1= list(outcomeOrdered= comfort1,
                     V= zeta_comfort1*comfort,
                     tau= list(tau_comfort1_1, tau_comfort1_2, tau_comfort1_3),
                     rows=comfort1!=0)
  ol_settings2= list(outcomeOrdered= comfort2,
                     V= zeta_comfort2*comfort,
                     tau= list(tau_comfort2_1, tau_comfort2_2, tau_comfort2_3, tau_comfort2_4),
                     rows=comfort2!=0)
  ol_settings3= list(outcomeOrdered= comfort3,
                     V= zeta_comfort3*comfort,
                     tau= list(tau_comfort3_1, tau_comfort3_2, tau_comfort3_3),
                     rows=comfort3!=0)
  ol_settings4= list(outcomeOrdered= comfort4,
                     V= zeta_comfort4*comfort,
                     tau= list(tau_comfort4_1, tau_comfort4_2, tau_comfort4_3),
                     rows=comfort4!=0)
  ol_settings5= list(outcomeOrdered= ttr1,
                     V= zeta_ttr1*ttr,
                     tau= list(tau_ttr1_1, tau_ttr1_2, tau_ttr1_3),
                     rows=ttr1!=0)
  ol_settings6= list(outcomeOrdered= ttr2,
                     V= zeta_ttr2*ttr,
                     tau= list(tau_ttr2_1, tau_ttr2_2, tau_ttr2_3),
                     rows=ttr2!=0)
  ol_settings7= list(outcomeOrdered= ttr3,
                     V= zeta_ttr3*ttr,
                     tau= list(tau_ttr3_1, tau_ttr3_2),
                     rows=ttr3!=0)
  ol_settings8= list(outcomeOrdered= cleanliness1,
                     V= zeta_cleanliness1*cleanliness,
                     tau= list(tau_cleanliness1_1, tau_cleanliness1_2),
                     rows=cleanliness1!=0)
  ol_settings9= list(outcomeOrdered= cleanliness2,
                     V= zeta_cleanliness2*cleanliness,
                     tau= list(tau_cleanliness2_1, tau_cleanliness2_2, tau_cleanliness2_3),
                     rows=cleanliness2!=0)
  ol_settings10= list(outcomeOrdered= staffbehaviour1,
                      V= zeta_staffbehaviour1*staffbehaviour,
                      tau= list(tau_staffbehaviour1_1, tau_staffbehaviour1_2, tau_staffbehaviour1_3, tau_staffbehaviour1_4),
                      rows=staffbehaviour1!=0)
  ol_settings11= list(outcomeOrdered= staffbehaviour2,
                      V= zeta_staffbehaviour2*staffbehaviour,
                      tau= list(tau_staffbehaviour2_1, tau_staffbehaviour2_2, tau_staffbehaviour2_3, tau_staffbehaviour2_4),
                      rows=staffbehaviour2!=0)
  ol_settings14= list(outcomeOrdered= envisusitain1,
                      V= zeta_envisusitain1*envisusitain,
                      tau= list(tau_envisusitain1_1, tau_envisusitain1_2),
                      rows=envisusitain1!=0)
  ol_settings15= list(outcomeOrdered= envisusitain2,
                      V= zeta_envisusitain2*envisusitain,
                      tau= list(tau_envisusitain2_1, tau_envisusitain2_2, tau_envisusitain2_3),
                      rows=envisusitain2!=0)
  
  
  P[["indic_comfort1"]]= apollo_ol(ol_settings1, functionality)
  P[["indic_comfort2"]]= apollo_ol(ol_settings2, functionality)
  P[["indic_comfort3"]]= apollo_ol(ol_settings3, functionality)
  P[["indic_comfort4"]]= apollo_ol(ol_settings4, functionality)
  P[["indic_ttr1"]]= apollo_ol(ol_settings5, functionality)
  P[["indic_ttr2"]]= apollo_ol(ol_settings6, functionality)
  P[["indic_ttr3"]]= apollo_ol(ol_settings7, functionality)
  P[["indic_cleanliness1"]]= apollo_ol(ol_settings8, functionality)
  P[["indic_cleanliness2"]]= apollo_ol(ol_settings9, functionality)
  P[["indic_staffbehaviour1"]]= apollo_ol(ol_settings10, functionality)
  P[["indic_staffbehaviour2"]]= apollo_ol(ol_settings11, functionality)
  P[["indic_envisusitain1"]]= apollo_ol(ol_settings14, functionality)
  P[["indic_envisusitain2"]]= apollo_ol(ol_settings15, functionality)
  
  
  ## Computing probabilities for MNL model component
  
  P[["choice"]]= apollo_mnl(mnl_settings, functionality)  
  
  ## Likelihood of the whole combined model
  
  P= apollo_combineModels(P, apollo_inputs, functionality)
  
  ### Average across inter-individual draws
  P = apollo_avgInterDraws(P, apollo_inputs, functionality)
  
  ## Prepare and return the outputs of the function
  
  P= apollo_prepareProb(P, apollo_inputs, functionality)
  
  return(P)
}

L <- apollo_probabilities(apollo_beta, apollo_inputs, 'estimate')									
## Model Estimation

# Likelihood before model estimation

apollo_llCalc(apollo_beta, apollo_probabilities, apollo_inputs)

# Estimate Model

model = apollo_estimate(apollo_beta, apollo_fixed, apollo_probabilities, apollo_inputs)

## Model Output

## apollo_modelOutput(model)

apollo_saveOutput(model)

I hope this time no generic error in code!
I can send the data to your email, if require.
Thank you in advance.

Post by **stephanehess** » 10 Jul 2024, 03:42

Hi

well, this is quite clear. You have included

Code: Select all

lambda_comfort*comfort + lambda_ttr*ttr + lambda_cleanliness*cleanliness + lambda_staffbehaviour*staffbehaviour + lambda_envisusitain*envisusitain

in both utilities.

Remember that only differences in utility matter, so these terms will just cancel out. You need to either interact them with an attribute that is alternative-specific or with the label of the alternatives (subject to a normalisation, so in this case only include them in one alternative)

Stephane

ApolloChoiceModelling forum

Error in ICLV estimation: "Estimation Failed. No covariance matrix to compute."

Error in ICLV estimation: "Estimation Failed. No covariance matrix to compute."

Re: Error in ICLV estimation: "Estimation Failed. No covariance matrix to compute."

Re: Error in ICLV estimation: "Estimation Failed. No covariance matrix to compute."

Re: Error in ICLV estimation: "Estimation Failed. No covariance matrix to compute."

Re: Error in ICLV estimation: "Estimation Failed. No covariance matrix to compute."

Re: Error in ICLV estimation: "Estimation Failed. No covariance matrix to compute."