When running on few features, SIVS throws error:
library(sivs)
# prepare toy data
data <- subset(iris, subset = Species != "setosa")
dim(data)
# 100 5
# run sivs
sivs_obj <- sivs::sivs(x = data[, -5],
y = factor(data[, 5]),
parallel.cores = 4,
progressbar = F)which produces
Error in apply(coef.df, 1, function(f) { :
dim(X) must have a positive length
In addition: Warning messages:
1: In `[<-.data.frame`(`*tmp*`, , tmp.columns.to.zscore, value = list( :
provided 400 variables to replace 4 variables
2: The method you have chosen is "glmnet" and it cannot handle missing values. The
provided data as 'x' argument contains 400 missing values. In total 100 rows
containing missing values were removed.
3: The method "glmnet" needs the argument 'family' to be filled. Since it was not
provided by the user and based on the factor levels of 'y' argument, the value
of argument 'family' has set to "binomial".
The error goes away if the parallel.cores is 3 or less, or even FALSE.
When running on few features, SIVS throws error:
which produces
The error goes away if the
parallel.coresis 3 or less, or evenFALSE.