Description
Hello,
I'am analysing sequences data for my master thesis. Doing so I'am intenting to use feature selection to find the most important features of my sequences to predict my outcome (salary at 30 years old).
When I run the following code in Rstudio I get an error:
Reglin_control_features_sala <- lm(t9salafteall_comp ~ Agglo + Sex + Migr_gen + Migr_reg + Langue_maison_dummy + Reg_ling, data=alldata_formation_salaires_NoNAs, weights = basewt)
set.seed(999)
y <- residuals(Reglin_control_features, type = "deviance")
x <- Features_12mois_SD
prop_stabsel <- stabs::stabsel( x= x , y = y, fitfun = glmnet.lasso, PFER=0.5, cutoff= 0.6,
B=1000)
=> "Error in res[[1]] : subscript out of bounds"
I tried different values for the parmeters PFER, cutoff and B, always with the same result, my professor, Matthias Studer, does not know where this error come from, that is why I am asking you directly.
Thanks a lot for your help,
Regards,
Leonhard.