looping through a list of independent variables

[adeldaoud]

Hi

Has anyone figured out how to loop/iterate through a list of independent variables (and dependent variables)? Preferably with the doParrallel package, as Chris suggested here https://www.cmm.bristol.ac.uk/forum/vie ... 43d352de0c

Suppose I want to iterate through the bang data with the following independent variable separately

Code: Select all

# (1) settings
library(doParallel)
library(R2MLwiN)
data(bang)

cl <- makeCluster(3)
registerDoParallel(cl) # register the cores


# this gives obviously an error since they are objects in the bang data. But I want them to eventually be part of the formula object defined below.
Focal_Independent_variables <- list(urban, age, hindu)


# (2) define the models to be passed to R2mlwin via doParallel
flist<-list(
  list(logit(use, cons) ~ 1 + Focal_Independent_variables  + (1 | district), D = "Binomial", estoptions = list(EstM = 0), data = bang))




result <- foreach(i=1:3, .packages="R2MLwiN", .errorhandling=c('pass')) %dopar% {
  do.call(runMLwiN, flist[[i]])
}

Essentially, the question is how can I combine various parts of a formula objects of R2mlwin?

[ChrisCharlton]

One possibility would be to build the required formulae as strings and then converting them into R formulae with the as.formula function (see https://stat.ethz.ch/R-manual/R-devel/l ... rmula.html). An example of this would be as follows:

Code: Select all

# (1) settings
library(doParallel)
library(R2MLwiN)
data(bang)

cl <- makeCluster(3)
registerDoParallel(cl) # register the cores


# this gives obviously an error since they are objects in the bang data. But I want them to eventually be part of the formula object defined below.
Focal_Independent_variables <- c("urban", "age", "hindu")


# (2) define the models to be passed to R2mlwin via doParallel
flist<-list()
for (i in 1:length(Focal_Independent_variables)) {
    flist[[i]] <- list(as.formula(paste0("logit(use, cons) ~ 1 +", Focal_Independent_variables[i], "+ (1 | district)")), D = "Binomial", estoptions = list(EstM = 0), data = bang)
}

result <- foreach::foreach(i=1:3, .packages="R2MLwiN", .errorhandling=c('pass')) %dopar% {
  do.call(runMLwiN, flist[[i]])
}

stopCluster(cl)

Code: Select all

> result
[[1]]

-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*- 
MLwiN (version: 2.36)  multilevel model (Binomial) 
          N min     mean max N_complete min_complete mean_complete max_complete
district 60   3 47.78333 173         60            3      47.78333          173
Estimation algorithm:  IGLS MQL1        Elapsed time : 0.14s 
Number of obs:  2867 (from total 2867)        The model converged after 5 iterations.
Log likelihood:      NA 
Deviance statistic:  NA 
--------------------------------------------------------------------------------------------------- 
The model formula:
logit(use, cons) ~ 1 + urban + (1 | district)
Level 2: district     Level 1: l1id      
--------------------------------------------------------------------------------------------------- 
The fixed part estimates:  
                Coef.   Std. Err.       z    Pr(>|z|)         [95% Conf.   Interval] 
Intercept    -0.65354     0.07607   -8.59   8.574e-18   ***     -0.80263    -0.50445 
urbanUrban    0.67291     0.09471    7.10   1.204e-12   ***      0.48728     0.85854 
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1  
--------------------------------------------------------------------------------------------------- 
The random part estimates at the district level: 
                  Coef.   Std. Err. 
var_Intercept   0.19017     0.05475 
--------------------------------------------------------------------------------------------------- 
The random part estimates at the l1id level: 
                Coef.   Std. Err. 
var_bcons_1   1.00000     0.00000 
-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*- 

[[2]]

-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*- 
MLwiN (version: 2.36)  multilevel model (Binomial) 
          N min     mean max N_complete min_complete mean_complete max_complete
district 60   3 47.78333 173         60            3      47.78333          173
Estimation algorithm:  IGLS MQL1        Elapsed time : 0.12s 
Number of obs:  2867 (from total 2867)        The model converged after 4 iterations.
Log likelihood:      NA 
Deviance statistic:  NA 
--------------------------------------------------------------------------------------------------- 
The model formula:
logit(use, cons) ~ 1 + age + (1 | district)
Level 2: district     Level 1: l1id      
--------------------------------------------------------------------------------------------------- 
The fixed part estimates:  
               Coef.   Std. Err.       z    Pr(>|z|)         [95% Conf.   Interval] 
Intercept   -0.47850     0.07741   -6.18   6.361e-10   ***     -0.63023    -0.32678 
age          0.01331     0.00429    3.10    0.001941   **       0.00489     0.02173 
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1  
--------------------------------------------------------------------------------------------------- 
The random part estimates at the district level: 
                  Coef.   Std. Err. 
var_Intercept   0.23992     0.06435 
--------------------------------------------------------------------------------------------------- 
The random part estimates at the l1id level: 
                Coef.   Std. Err. 
var_bcons_1   1.00000     0.00000 
-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*- 

[[3]]

-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*- 
MLwiN (version: 2.36)  multilevel model (Binomial) 
          N min     mean max N_complete min_complete mean_complete max_complete
district 60   3 47.78333 173         60            3      47.78333          173
Estimation algorithm:  IGLS MQL1        Elapsed time : 0.14s 
Number of obs:  2867 (from total 2867)        The model converged after 5 iterations.
Log likelihood:      NA 
Deviance statistic:  NA 
--------------------------------------------------------------------------------------------------- 
The model formula:
logit(use, cons) ~ 1 + hindu + (1 | district)
Level 2: district     Level 1: l1id      
--------------------------------------------------------------------------------------------------- 
The fixed part estimates:  
                Coef.   Std. Err.       z    Pr(>|z|)         [95% Conf.   Interval] 
Intercept    -0.53815     0.07835   -6.87   6.484e-12   ***     -0.69171    -0.38459 
hinduHindu    0.39805     0.11845    3.36   0.0007779   ***      0.16590     0.63021 
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1  
--------------------------------------------------------------------------------------------------- 
The random part estimates at the district level: 
                  Coef.   Std. Err. 
var_Intercept   0.23016     0.06220 
--------------------------------------------------------------------------------------------------- 
The random part estimates at the l1id level: 
                Coef.   Std. Err. 
var_bcons_1   1.00000     0.00000 
-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-

[adeldaoud]

Sorry for the late reply. This works perfectly.

Thanks Chris