ApolloChoiceModelling forum

Hi Stephane or David,

I wonder why I keep getting NAs in the hybrid model. It worked when I start with a simpler one, with one or two latent variables (FNS1, FNS2). But when I include more, the model results NA for estimates, std error, and so on.

My code are as below:

# ################################################################# #
#### LOAD LIBRARY AND DEFINE CORE SETTINGS ####
# ################################################################# #

### Clear memory
rm(list = ls())

### Load Apollo library
library(apollo)

### Initialise code
apollo_initialise()

### Set core controls
apollo_control = list(
modelName = "rice_Hybrid_with_OL",
modelDescr = "Hybrid choice model on rice choice data, using ordered measurement model for indicators",
indivID = "id",
nCores = 10
)

# ################################################################# #
#### LOAD DATA AND APPLY ANY TRANSFORMATIONS ####
# ################################################################# #

### Loading data from package
### if data is to be loaded from a file (e.g. called data.csv),
### the code would be: database = read.csv("data.csv",header=TRUE)
database = read.csv("D:/2023/3、院长基金/0722 main data/ricedata_main.csv",header=TRUE)

# ################################################################# #
#### ANALYSIS OF CHOICES ####
# ################################################################# #

### Illustration of how to use apollo_choiceAnalysis with user-defined alternatives.
### This is useful in cases where the alternatives in the data differ
### across tasks. The same approach can then also be used with unlabelled data

choiceAnalysis_settings <- list(
alternatives = c(option_1=1, option_2=2, option_3=3, option_4=4),
avail = 1,
choiceVar = database$choice,
explanators = database[,c("edu","age","marry","income","gender")]
)

apollo_choiceAnalysis(choiceAnalysis_settings, apollo_control, database)

# ################################################################# #
#### DEFINE MODEL PARAMETERS ####
# ################################################################# #

### Vector of parameters, including any that are kept fixed in estimation
apollo_beta = c(asc_non_purchase = -1.52,
b_white = -0.05,
b_brown = -0.17,
b_organic = 0.57,
b_brand = 0.13,
b_lowGI = 0.99,
b_lowgluten = 0.98,
b_enhanced = 1.38,
b_price = -0.06,
lambda = 1,
gamma_ricehabit = 0,
gamma_younger = 0,
gamma_bachelor = 0,
gamma_kid = 0,
gamma_highincome = 0,
gamma_gender = 0,
zeta_FNS1 = 0,
zeta_FNS2 = 0,
zeta_FNS3 = 0,
zeta_FNS4 = 0,
zeta_FNS5 = 0,
zeta_FNS6 = 0,
zeta_FNS7 = 0,
zeta_FNS8 = 0,
zeta_FNS9 = 0,
tau_FNS1_1 = -3,
tau_FNS1_2 = -2,
tau_FNS1_3 = -1,
tau_FNS1_4 = 1,
tau_FNS1_5 = 2,
tau_FNS1_6 = 3,
tau_FNS2_1 = -3,
tau_FNS2_2 = -2,
tau_FNS2_3 = -1,
tau_FNS2_4 = 1,
tau_FNS2_5 = 2,
tau_FNS2_6 = 3,
tau_FNS3_1 = -3,
tau_FNS3_2 = -2,
tau_FNS3_3 = -1,
tau_FNS3_4 = 1,
tau_FNS3_5 = 2,
tau_FNS3_6 = 3,
tau_FNS4_1 = -3,
tau_FNS4_2 = -2,
tau_FNS4_3 = -1,
tau_FNS4_4 = 1,
tau_FNS4_5 = 2,
tau_FNS4_6 = 3,
tau_FNS5_1 = -3,
tau_FNS5_2 = -2,
tau_FNS5_3 = -1,
tau_FNS5_4 = 1,
tau_FNS5_5 = 2,
tau_FNS5_6 = 3,
tau_FNS6_1 = -3,
tau_FNS6_2 = -2,
tau_FNS6_3 = -1,
tau_FNS6_4 = 1,
tau_FNS6_5 = 2,
tau_FNS6_6 = 3,
tau_FNS7_1 = -3,
tau_FNS7_2 = -2,
tau_FNS7_3 = -1,
tau_FNS7_4 = 1,
tau_FNS7_5 = 2,
tau_FNS7_6 = 3,
tau_FNS8_1 = -3,
tau_FNS8_2 = -2,
tau_FNS8_3 = -1,
tau_FNS8_4 = 1,
tau_FNS8_5 = 2,
tau_FNS8_6 = 3,
tau_FNS9_1 = -3,
tau_FNS9_2 = -2,
tau_FNS9_3 = -1,
tau_FNS9_4 = 1,
tau_FNS9_5 = 2,
tau_FNS9_6 = 3
)
### 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()

# ################################################################# #
#### DEFINE RANDOM COMPONENTS ####
# ################################################################# #

### Set parameters for generating draws
apollo_draws = list(
interDrawsType="halton",
interNDraws=100,
interUnifDraws=c(),
interNormDraws=c("eta"),

intraDrawsType="",
intraNDraws=0,
intraUnifDraws=c(),
intraNormDraws=c()
)

### Create random parameters
apollo_randCoeff=function(apollo_beta, apollo_inputs){
randcoeff = list()

randcoeff[["LV"]] = gamma_ricehabit * (ricehabit>3) + gamma_younger * (age<2) + gamma_bachelor * edu + gamma_kid * (marry==2|marry==4) + gamma_highincome * (income>1) + gamma_gender * gender + eta

return(randcoeff)
}

# ################################################################# #
#### GROUP AND VALIDATE INPUTS ####
# ################################################################# #

apollo_inputs = apollo_validateInputs()

# ################################################################# #
#### DEFINE MODEL 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 list of probabilities P
P = list()

### Likelihood of indicators
ol_settings1 = list(outcomeOrdered = FNS1,
V = zeta_FNS1*LV,
tau = list(tau_FNS1_1, tau_FNS1_2, tau_FNS1_3, tau_FNS1_4, tau_FNS1_5, tau_FNS1_6),
rows = (task==1|task==2),
componentName = "indic_FNS1")
ol_settings2 = list(outcomeOrdered = FNS1,
V = zeta_FNS2*LV,
tau = list(tau_FNS2_1, tau_FNS2_2, tau_FNS2_3, tau_FNS2_4, tau_FNS2_5, tau_FNS2_6),
rows = (task==1|task==2),
componentName = "indic_FNS2")
ol_settings3 = list(outcomeOrdered = FNS3,
V = zeta_FNS3*LV,
tau = list(tau_FNS3_1, tau_FNS3_2, tau_FNS3_3, tau_FNS3_4, tau_FNS3_5, tau_FNS3_6),
rows = (task==1|task==2),
componentName = "indic_FNS3")
ol_settings4 = list(outcomeOrdered = FNS4,
V = zeta_FNS4*LV,
tau = list(tau_FNS4_1, tau_FNS4_2, tau_FNS4_3, tau_FNS4_4, tau_FNS4_5, tau_FNS4_6),
rows = (task==1|task==2),
componentName = "indic_FNS4")
ol_settings5 = list(outcomeOrdered = FNS5,
V = zeta_FNS5*LV,
tau = list(tau_FNS5_1, tau_FNS5_2, tau_FNS5_3, tau_FNS5_4, tau_FNS5_5, tau_FNS5_6),
rows = (task==1|task==2),
componentName = "indic_FNS5")
ol_settings6 = list(outcomeOrdered = FNS6,
V = zeta_FNS6*LV,
tau = list(tau_FNS6_1, tau_FNS6_2, tau_FNS6_3, tau_FNS6_4, tau_FNS6_5, tau_FNS6_6),
rows = (task==1|task==2),
componentName = "indic_FNS6")
ol_settings7 = list(outcomeOrdered = FNS7,
V = zeta_FNS7*LV,
tau = list(tau_FNS7_1, tau_FNS7_2, tau_FNS7_3, tau_FNS7_4, tau_FNS7_5, tau_FNS7_6),
rows = (task==1|task==2),
componentName = "indic_FNS7")
ol_settings8 = list(outcomeOrdered = FNS8,
V = zeta_FNS8*LV,
tau = list(tau_FNS8_1, tau_FNS8_2, tau_FNS8_3, tau_FNS8_4, tau_FNS8_5, tau_FNS8_6),
rows = (task==1|task==2),
componentName = "indic_FNS8")
ol_settings9 = list(outcomeOrdered = FNS9,
V = zeta_FNS9*LV,
tau = list(tau_FNS9_1, tau_FNS9_2, tau_FNS9_3, tau_FNS9_4, tau_FNS9_5, tau_FNS9_6),
rows = (task==1|task==2),
componentName = "indic_FNS9")
P[["indic_FNS1"]] = apollo_ol(ol_settings1, functionality)
P[["indic_FNS2"]] = apollo_ol(ol_settings2, functionality)
P[["indic_FNS3"]] = apollo_ol(ol_settings3, functionality)
P[["indic_FNS4"]] = apollo_ol(ol_settings4, functionality)
P[["indic_FNS5"]] = apollo_ol(ol_settings5, functionality)
P[["indic_FNS6"]] = apollo_ol(ol_settings6, functionality)
P[["indic_FNS7"]] = apollo_ol(ol_settings7, functionality)
P[["indic_FNS8"]] = apollo_ol(ol_settings8, functionality)
P[["indic_FNS9"]] = apollo_ol(ol_settings9, functionality)

### Likelihood of choices
### List of utilities: these must use the same names as in mnl_settings, order is irrelevant
V = list()
V[['option_1']] = b_white * (white_1==0) + b_brown * (white_1==2) + b_organic * organic_1 + b_brand * brand_1 + b_lowGI * (health_1==1) + b_lowgluten * (health_1==2) + b_enhanced * (health_1==3) + b_price * price_1 + lambda * LV
V[['option_2']] = b_white * (white_2==0) + b_brown * (white_2==2) + b_organic * organic_2 + b_brand * brand_2 + b_lowGI * (health_2==1) + b_lowgluten * (health_2==2) + b_enhanced * (health_2==3) + b_price * price_2 + lambda * LV
V[['option_3']] = b_white * (white_3==0) + b_brown * (white_3==2) + b_organic * organic_3 + b_brand * brand_3 + b_lowGI * (health_3==1) + b_lowgluten * (health_3==2) + b_enhanced * (health_3==3) + b_price * price_3 + lambda * LV
V[['option_4']] = asc_non_purchase

### Define settings for MNL model component
mnl_settings = list(
alternatives = c('option_1'=1, 'option_2'=2, 'option_3'=3, 'option_4'=4),
avail = 1,
choiceVar = choice,
utilities = V,
componentName = "choice"
)

### Compute probabilities for MNL model component
P[["choice"]] = apollo_mnl(mnl_settings, functionality)

### Likelihood of the whole model
P = apollo_combineModels(P, apollo_inputs, functionality)

### Take product across observation for same individual
P = apollo_panelProd(P, apollo_inputs, functionality)

### Average across inter-individual draws
P = apollo_avgInterDraws(P, apollo_inputs, functionality)

### Prepare and return outputs of function
P = apollo_prepareProb(P, apollo_inputs, functionality)
return(P)
}

# ################################################################# #
#### MODEL ESTIMATION ####
# ################################################################# #

### Optional: calculate LL 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 OUTPUTS ####
# ################################################################# #

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

apollo_modelOutput(model)

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

apollo_saveOutput(model)

Could you please suggest what to do to solve the problem?

Best,
Xian

Hi

can you show us the results? There could be many reasons.

Also, please note that your model only has 1 latent variable (LV). What you are referring to are the indicators, not the latent variables

Stephane

Dear Stephane,

I'm sorry that this post should be deleted because I have asked you the same question in another part of the forum.

Thank you pointing out my misunderstanding about latent variable.

Best,
Xian