0

I am trying to apply logistf() on a dataframe (dim:11359x139). All the variables are binary. I get the following message: "Error in logistf.fit(x = x, y = y, weight = weight, offset = offset, firth, : no memory available". Even if I take into consideration only 20 rows and 139 predictors of the dataframe I get the same. Is it a hardware issue or my fault?

elpavlos
  • 35
  • 4
  • Try less predictors. You should have (much) more samples (rows) than variables to have anything meaningful. Otherwise I can tell you that no variable contributes anything to your model without running the code at all. – Roman Luštrik Jun 15 '17 at 18:20
  • Thanks! I'll give it another try with less predictors. Is there a way to estimate how much more samples than variables you need in order to get something meaningful? – elpavlos Jun 15 '17 at 20:56
  • https://journals.sagepub.com/doi/10.1177/0962280218784726, this might be helpful – Roasty247 Apr 02 '19 at 12:12

0 Answers0