CODEWITHSUNDEEP
rfeature-extractionfselector
Ranjana Girish
Ranjana Girish

Reputation: 473

Regarding usage of sparse matrix as an input in Fselector package

I am using information.gain(), chi.squared() and gain. Ratio() functions provided by Fselector package for feature selection.

All these functions are accepting data frame as an input, as I am using matrix of dimension 200161(rows), 323(Column) data set ,Getting error 

error in jnew(name): java.lang.outofMemoryError: java heap space R

I have xeon preocessor, with 8 GB RAM, and after expanding the memory limit to max to assert by using command 

memory.limit()

results in 8026,which is also not sufficient

For memory efficiency i have converted data frame to Sparse matrix

Now the question is, can we use sparse matrix as an input to functions information.gain(), chi.squared() and gain.ratio()

Please anyone help regarding this,

Thanks in advance

Upvotes: 0

Views: 72

Answers (1)

Lars Kotthoff
Lars Kotthoff

Reputation: 109242

This is not supported at the moment, sorry.

Upvotes: 0

Related Questions