Error: the variable does not influence the log-likelihood of your model

[oleggurshev]

Hello professor,

We have been using your package in our classes (University of Warsaw) for micromodelling. We were modelling society choice (1 - society1, 2-society2, 3- opt-out of the survey). I have run into an issue where the function would not recognize any of my variables inside a utility (Error in apollo_estimate(apollo_beta, apollo_fixed, apollo_probabilities, : Parameter asc_opt1 does not influence the log-likelihood of your model!).

I am including both data and my code.

### Load data
setwd("")
database <- read_xlsx('Data_relative.xlsx')

database <- transmute(database, #modify this if want logs / not logs
ID = ID,
income1 = log(income1),
income2 = log(income2),
choice = choice,
choice1 = choice1,
choice2 = choice2,
choice3 = choice3,
centile1 = log(centile1),
centile2 = log(centile2),
trend1 = trend1,
trend2 = trend2
)

database <- as.data.frame(database) #apollo does not work with tibbles

### Initialize code
apollo_initialise()

### Set core controls
apollo_control = list(
modelName ="Society_base",
modelDescr ="Simple MNL model on society choice data",
indivID ="ID"
)

# ################################################################# #
#### ANALYSIS OF CHOICES ####
# ################################################################# #
database$low_income1=database$income1<quantile(database$income1,0.25)
database$low_income2=database$income2<quantile(database$income2,0.25)


df <- database %>% select(income1:income2) #modify if want different explanators


choiceAnalysis_settings <- list(
alternatives = c(opt1 = 1, opt2 = 2, out = 3),
avail = with(database, list(
opt1 = choice1, opt2 = choice2, out = choice3
)),
choiceVar = database$choice,
explanators = database
)


apollo_choiceAnalysis(choiceAnalysis_settings, apollo_control, database) #this does not work, Error in if (x[1] == x[2]) { : missing value where TRUE/FALSE needed

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

### Vector of parameters, including any that are kept fixed in estimation
apollo_beta = c(
asc_opt1 = 0,
asc_opt2 = 0,
asc_out =0,
b_income = 0,
b_trend = 0,
b_centile = 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_out')

# ################################################################# #
#### 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()

### List of utilities: these must use the same names as in mnl_settings, order is irrelevant
V = list()
V[['out']] = asc_out
V[['opt1']] = asc_opt1 + b_income * income1 + b_trend * trend1 + b_centile * centile1
V[['opt2']] = asc_opt2 + b_income * income2 + b_trend * trend2 + b_centile * centile2


### Define settings for MNL model component
mnl_settings = list(
alternatives = c(opt1=1, opt2=2, out=3),
avail = list(opt1=choice1, opt2 = choice2, out=choice3),
choiceVar = choice,
V = V
)

### Compute probabilities using MNL model
P[['model']] = apollo_mnl(mnl_settings, functionality)

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

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

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

[stephanehess]

Hi

the data was not included, so we can't test this ourselves, but one suspicion is that it could be caused by loading the data from an Excel file. We recommend working with csv. If that doesn't solve it, then please send us the data and we'll have a look

Thanks

Stephane