Error: Non-valid user-provided bgw_setting setting name detected: "writeIterMode"."
Posted: 14 Aug 2023, 10:58
Dear Prof. Hess,
I have previously ran successfully, an MNL model on Apollo version v0.2.9 (about 6 weeks ago) which I used to obtain priors. I have now come back with full data and run the same model again (albeit different data).During this, I was prompted to update to Apollo v0.3.0 but since updating, during model estimation, I receive the following error:
Error in bgw::bgw_mle(calcR = f, betaStart = beta_var_val, calcJ = grad, :
Non-valid user-provided bgw_setting setting name detected: "writeIterMode"."
I cannot find anything related to bgw_setting in the R script or manual. The only thing I can see is that it writes the estimated parameter values in each iteration to a file in the working/output directory so the problem might be associated with this?
Any help would be very much appreciated.
Script below:
Output below:
I have previously ran successfully, an MNL model on Apollo version v0.2.9 (about 6 weeks ago) which I used to obtain priors. I have now come back with full data and run the same model again (albeit different data).During this, I was prompted to update to Apollo v0.3.0 but since updating, during model estimation, I receive the following error:
Error in bgw::bgw_mle(calcR = f, betaStart = beta_var_val, calcJ = grad, :
Non-valid user-provided bgw_setting setting name detected: "writeIterMode"."
I cannot find anything related to bgw_setting in the R script or manual. The only thing I can see is that it writes the estimated parameter values in each iteration to a file in the working/output directory so the problem might be associated with this?
Any help would be very much appreciated.
Script below:
Code: Select all
# ### ############################################################# #
#### LOAD LIBRARY AND DEFINE CORE SETTINGS ####
# ################################################################# #
### Clear memory
rm(list = ls())
### Load libraries
library(apollo)
### Initialise code
apollo_initialise()
### Set core controls
apollo_control = list(
modelName = "MNL_Nonspeeders_all_data_BASE",
modelDescr = "First MNL model using Nonspeeders_all_data",
indivID = "RID",
outputDirectory = "MNL_Nonspeeders_all_data_BASE_OUTPUT"
)
# ################################################################# #
#### LOAD DATA AND APPLY ANY TRANSFORMATIONS ####
# ################################################################ #
database = read.csv("Nonspeeders_all_data.csv",header=TRUE)
### Run function to analyse choice data
#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:
asc_DrugA = 0,
#main effects:
b_CERT_very_low = 0,
b_CERT_low = 0,
b_CERT_moderate = 0,
b_CERT_high = 0,
b_WAIT = 0,
#interaction effects (with WAIT) for scenario variables:
b_LIFExWAIT = 0,
b_QOL_light_workxWAIT = 0,
# b_QOL_no_workxWAIT = 0,
# b_QOL_bed_50xWAIT = 0,
b_QOL_comp_disabledxWAIT = 0,
b_BENEFIT_smallxWAIT = 0,
# b_BENEFIT_moderatexWAIT = 0,
b_BENEFIT_substantialxWAIT= 0,
#interaction effects
#(with CERT) for
#scenario variables:
# b_LIFExCERT_very_low = 0,
# b_LIFExCERT_low = 0,
# b_LIFExCERT_moderate = 0,
# b_LIFExCERT_high = 0,
b_QOL_light_workxCERT_very_low = 0,
# b_QOL_no_workxCERT_very_low = 0,
# b_QOL_bed_50xCERT_very_low = 0,
b_QOL_comp_disabledxCERT_very_low = 0,
# b_QOL_light_workxCERT_low = 0,
# b_QOL_no_workxCERT_low = 0,
# b_QOL_bed_50xCERT_low = 0,
# b_QOL_comp_disabledxCERT_low = 0,
# b_QOL_light_workxCERT_moderate = 0,
# b_QOL_no_workxCERT_moderate = 0,
# b_QOL_bed_50xCERT_moderate = 0,
# b_QOL_comp_disabledxCERT_moderate = 0,
b_QOL_light_workxCERT_high = 0,
# b_QOL_no_workxCERT_high = 0,
# b_QOL_bed_50xCERT_high = 0,
b_QOL_comp_disabledxCERT_high = 0,
b_BENEFIT_smallxCERT_very_low = 0,
# b_BENEFIT_moderatexCERT_very_low = 0,
b_BENEFIT_substantialxCERT_very_low = 0,
# b_BENEFIT_smallxCERT_low = 0,
# b_BENEFIT_moderatexCERT_low = 0,
# b_BENEFIT_substantialxCERT_low = 0,
# b_BENEFIT_smallxCERT_moderate = 0,
# b_BENEFIT_moderatexCERT_moderate = 0,
# b_BENEFIT_substantialxCERT_moderate = 0,
b_BENEFIT_smallxCERT_high = 0,
# b_BENEFIT_moderatexCERT_high = 0,
b_BENEFIT_substantialxCERT_high = 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_DrugA",
"b_CERT_very_low")
# "b_QOL_light_workxWAIT",
# "b_BENEFIT_smallxWAIT",
# "b_LIFExCERT_very_low",
# "b_QOL_light_workxCERT_very_low",
# "b_QOL_no_workxCERT_very_low",
# "b_QOL_bed_50xCERT_very_low",
# "b_QOL_comp_disabledxCERT_very_low",
# "b_BENEFIT_smallxCERT_very_low",
# "b_BENEFIT_moderatexCERT_very_low",
# "b_BENEFIT_substantialxCERT_very_low")
# ################################################################# #
#### 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[["DrugA"]] =
asc_DrugA +
b_CERT_very_low * (DrugA_CERT == 1) +
b_CERT_low * (DrugA_CERT == 2) +
b_CERT_moderate * (DrugA_CERT == 3) +
b_CERT_high * (DrugA_CERT == 4) +
b_WAIT * DrugA_WAIT +
#Interactions with WAIT:
b_LIFExWAIT * DrugA_WAIT * LIFE +
b_QOL_light_workxWAIT * ( DrugA_WAIT * (QOL == 1)) +
# b_QOL_no_workxWAIT * ( DrugA_WAIT * (QOL == 2)) +
# b_QOL_bed_50xWAIT * ( DrugA_WAIT * (QOL == 3)) +
b_QOL_comp_disabledxWAIT * ( DrugA_WAIT * (QOL == 4)) +
b_BENEFIT_smallxWAIT * ( DrugA_WAIT * (BENEFIT == 1)) +
# b_BENEFIT_moderatexWAIT * ( DrugA_WAIT * (BENEFIT == 2)) +
b_BENEFIT_substantialxWAIT * ( DrugA_WAIT * (BENEFIT == 3)) +
#Interactions with CERT (all categorical levels):
# b_LIFExCERT_very_low * ( DrugA_CERT == 1) * LIFE +
# b_LIFExCERT_low * ( DrugA_CERT == 2) * LIFE +
# b_LIFExCERT_moderate * ( DrugA_CERT == 3) * LIFE +
# b_LIFExCERT_high * ( DrugA_CERT == 4) * LIFE +
b_QOL_light_workxCERT_very_low * ( (DrugA_CERT == 1) * (QOL == 1)) +
# b_QOL_light_workxCERT_low * ( (DrugA_CERT == 2) * (QOL == 1)) +
# b_QOL_light_workxCERT_moderate * ( (DrugA_CERT == 3) * (QOL == 1)) +
b_QOL_light_workxCERT_high * ( (DrugA_CERT == 4) * (QOL == 1)) +
# b_QOL_no_workxCERT_very_low * ( (DrugA_CERT == 1) * (QOL == 2)) +
# b_QOL_no_workxCERT_low * ( (DrugA_CERT == 2) * (QOL == 2)) +
# b_QOL_no_workxCERT_moderate * ( (DrugA_CERT == 3) * (QOL == 2)) +
# b_QOL_no_workxCERT_high * ( (DrugA_CERT == 4) * (QOL == 2)) +
# b_QOL_bed_50xCERT_very_low * ( (DrugA_CERT == 1) * (QOL == 3)) +
# b_QOL_bed_50xCERT_low * ( (DrugA_CERT == 2) * (QOL == 3)) +
# b_QOL_bed_50xCERT_moderate * ( (DrugA_CERT == 3) * (QOL == 3)) +
# b_QOL_bed_50xCERT_high * ( (DrugA_CERT == 4) * (QOL == 3)) +
b_QOL_comp_disabledxCERT_very_low * ( (DrugA_CERT == 1) * (QOL == 4)) +
# b_QOL_comp_disabledxCERT_low * ( (DrugA_CERT == 2) * (QOL == 4)) +
# b_QOL_comp_disabledxCERT_moderate * ( (DrugA_CERT == 3) * (QOL == 4)) +
b_QOL_comp_disabledxCERT_high * ( (DrugA_CERT == 4) * (QOL == 4)) +
b_BENEFIT_smallxCERT_very_low * ( (DrugA_CERT == 1) * (BENEFIT == 1)) +
# b_BENEFIT_smallxCERT_low * ( (DrugA_CERT == 2) * (BENEFIT == 1)) +
# b_BENEFIT_smallxCERT_moderate * ( (DrugA_CERT == 3) * (BENEFIT == 1)) +
b_BENEFIT_smallxCERT_high * ( (DrugA_CERT == 4) * (BENEFIT == 1)) +
# b_BENEFIT_moderatexCERT_very_low * ( (DrugA_CERT == 1) * (BENEFIT == 2)) +
# b_BENEFIT_moderatexCERT_low * ( (DrugA_CERT == 2) * (BENEFIT == 2)) +
# b_BENEFIT_moderatexCERT_moderate * ( (DrugA_CERT == 3) * (BENEFIT == 2)) +
# b_BENEFIT_moderatexCERT_high * ( (DrugA_CERT == 4) * (BENEFIT == 2)) +
b_BENEFIT_substantialxCERT_very_low * ( (DrugA_CERT == 1) * (BENEFIT == 3)) +
# b_BENEFIT_substantialxCERT_low * ( (DrugA_CERT == 2) * (BENEFIT == 3)) +
# b_BENEFIT_substantialxCERT_moderate * ( (DrugA_CERT == 3) * (BENEFIT == 3)) +
b_BENEFIT_substantialxCERT_high * ( (DrugA_CERT == 4) * (BENEFIT == 3))
V[["DrugB"]] =
b_CERT_very_low * (DrugB_CERT == 1) +
b_CERT_low * (DrugB_CERT == 2) +
b_CERT_moderate * (DrugB_CERT == 3) +
b_CERT_high * (DrugB_CERT == 4) +
b_WAIT * DrugB_WAIT +
#Interactions with WAIT:
b_LIFExWAIT * DrugB_WAIT * LIFE +
b_QOL_light_workxWAIT * ( DrugB_WAIT * (QOL == 1)) +
# b_QOL_no_workxWAIT * ( DrugB_WAIT * (QOL == 2)) +
# b_QOL_bed_50xWAIT * ( DrugB_WAIT * (QOL == 3)) +
b_QOL_comp_disabledxWAIT * ( DrugB_WAIT * (QOL == 4)) +
b_BENEFIT_smallxWAIT * ( DrugB_WAIT * (BENEFIT == 1)) +
# b_BENEFIT_moderatexWAIT * ( DrugB_WAIT * (BENEFIT == 2)) +
b_BENEFIT_substantialxWAIT * ( DrugB_WAIT * (BENEFIT == 3)) +
#Interactions with CERT (all categorical levels):
# b_LIFExCERT_very_low * (DrugB_CERT == 1) * LIFE +
# b_LIFExCERT_low * (DrugB_CERT == 2) * LIFE +
# b_LIFExCERT_moderate * (DrugB_CERT == 3) * LIFE +
# b_LIFExCERT_high * (DrugB_CERT == 4) * LIFE
b_QOL_light_workxCERT_very_low * ( (DrugB_CERT == 1) * (QOL == 1)) +
# b_QOL_light_workxCERT_low * ( (DrugB_CERT == 2) * (QOL == 1)) +
# b_QOL_light_workxCERT_moderate * ( (DrugB_CERT == 3) * (QOL == 1)) +
b_QOL_light_workxCERT_high * ( (DrugB_CERT == 4) * (QOL == 1)) +
# b_QOL_no_workxCERT_very_low * ( (DrugB_CERT == 1) * (QOL == 2)) +
# b_QOL_no_workxCERT_low * ( (DrugB_CERT == 2) * (QOL == 2)) +
# b_QOL_no_workxCERT_moderate * ( (DrugB_CERT == 3) * (QOL == 2)) +
# b_QOL_no_workxCERT_high * ( (DrugB_CERT == 4) * (QOL == 2)) +
# b_QOL_bed_50xCERT_very_low * ( (DrugB_CERT == 1) * (QOL == 3)) +
# b_QOL_bed_50xCERT_low * ( (DrugB_CERT == 2) * (QOL == 3)) +
# b_QOL_bed_50xCERT_moderate * ( (DrugB_CERT == 3) * (QOL == 3)) +
# b_QOL_bed_50xCERT_high * ( (DrugB_CERT == 4) * (QOL == 3)) +
b_QOL_comp_disabledxCERT_very_low * ( (DrugB_CERT == 1) * (QOL == 4)) +
# b_QOL_comp_disabledxCERT_low * ( (DrugB_CERT == 2) * (QOL == 4)) +
# b_QOL_comp_disabledxCERT_moderate * ( (DrugB_CERT == 3) * (QOL == 4)) +
b_QOL_comp_disabledxCERT_high * ( (DrugB_CERT == 4) * (QOL == 4)) +
b_BENEFIT_smallxCERT_very_low * ( (DrugB_CERT == 1) * (BENEFIT == 1)) +
# b_BENEFIT_smallxCERT_low * ( (DrugB_CERT == 2) * (BENEFIT == 1)) +
# b_BENEFIT_smallxCERT_moderate * ( (DrugB_CERT == 3) * (BENEFIT == 1)) +
b_BENEFIT_smallxCERT_high * ( (DrugB_CERT == 4) * (BENEFIT == 1)) +
# b_BENEFIT_moderatexCERT_very_low * ( (DrugB_CERT == 1) * (BENEFIT == 2)) +
# b_BENEFIT_moderatexCERT_low * ( (DrugB_CERT == 2) * (BENEFIT == 2)) +
# b_BENEFIT_moderatexCERT_moderate * ( (DrugB_CERT == 3) * (BENEFIT == 2)) +
# b_BENEFIT_moderatexCERT_high * ( (DrugB_CERT == 4) * (BENEFIT == 2)) +
b_BENEFIT_substantialxCERT_very_low * ( (DrugB_CERT == 1) * (BENEFIT == 3)) +
# b_BENEFIT_substantialxCERT_low * ( (DrugB_CERT == 2) * (BENEFIT == 3)) +
# b_BENEFIT_substantialxCERT_moderate * ( (DrugB_CERT == 3) * (BENEFIT == 3)) +
b_BENEFIT_substantialxCERT_high * ( (DrugB_CERT == 4) * (BENEFIT == 3))
### Define settings for MNL model component
mnl_settings = list(
alternatives = c(DrugA=1, DrugB=2),
choiceVar = CHOICE,
utilities = 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)
}
apollo_probabilities(apollo_beta, apollo_inputs, functionality="estimate")
# ################################################################# #
#### MODEL ESTIMATION ####
# ################################################################# #
model = apollo_estimate(apollo_beta, apollo_fixed, apollo_probabilities, apollo_inputs)
# ################################################################# #
#### MODEL OUTPUTS ####
# ################################################################# #
# ----------------------------------------------------------------- #
#---- FORMATTED OUTPUT (TO SCREEN) ----
# ----------------------------------------------------------------- #
apollo_modelOutput(model)
Code: Select all
> # ################################################################# #
> #### MODEL ESTIMATION ####
> # ################################################################# #
>
> model = apollo_estimate(apollo_beta, apollo_fixed, apollo_probabilities, apollo_inputs)
Preparing user-defined functions.
Testing likelihood function...
Setting "avail" is missing, so full availability is assumed.
Overview of choices for MNL model component :
DrugA DrugB
Times available 10440.00 10440.00
Times chosen 5621.00 4819.00
Percentage chosen overall 53.84 46.16
Percentage chosen when available 53.84 46.16
Pre-processing likelihood function...
Testing influence of parameters
Starting main estimation
BGW using analytic model derivatives supplied by caller...
Error in bgw::bgw_mle(calcR = f, betaStart = beta_var_val, calcJ = grad, :
Non-valid user-provided bgw_setting setting name detected: "writeIterMode".