[R] Active learning and logistic regression.

Łukasz Lew lukasz.lew at gmail.com
Sun Oct 11 15:24:22 CEST 2009


I have binary file with a set of tunable real parameters.
Upon execution the file returns 0 or 1 with some probability.
(therefore (nonlinear) logistic regression)
I need to find a set of parameters for which the probability of 1 is high.

I would like to create program that would repeatably execute my file
with different parameters,
get 0,1 response after each execution (therefore active learning)
and build a model of the underlying probability distribution.

How to aproach this problem with R?
Are there libraries, scripts that I could start with?

Lukasz Lew

More information about the R-help mailing list