What I want to do:
I plan to do a logistic regression with several independant variables. I have been using excel for several linear regression models to this point. I tried to get a logistic setup in excel by telling the excel solver to maximize the total probability, but the result was not accurate percentages, as there were 6 out of 97 predictions of 99+ percent that were incorrect.
I am wondering what you guys' thoughts are on my options for the software I use. I have heard good things about Octave, R, and gretl, but I have used none of them and am even unsure if they can do what I need them to do period, much less accurately.
FTR, I am an experienced programmer (C, C++, Java, Python), and should not have much trouble learning any of the programs.
Out of the options you listed I only have experience with R. There's some what of a learning curve associated with it but since you're a programmer I'm not too worried about that. R will be able to do what you want it to do.
I have heard a lot of good things about gretl. has anyone used it?
Advertise on Talk Stats