Logistic Regression only for 2 Label Classes?

Fred12Fred12 MemberPosts:344Unicorn
edited November 2019 inHelp

hi,
I had some issues with logistic regression.. to clarify first: can it only handle binomial values, that's said, only label class with 2 possible values ? or also more? because I had nominal label data with 3 classes... it says "cannot handle polynomial label"... when I splitted the dataset into only 2 possible label class values, it seemed to run at least, but then my Rapidminer Studio crashes every time I run it... is it a bug?

Tagged:

Answers

  • zprekopcsakzprekopcsak RapidMiner Certified Expert, MemberPosts:47Guru

    Fred,

    Can you please share your crashing process and some sample of the data?

    Thanks, Zoltan

  • sfraunhoffersfraunhoffer MemberPosts:4Contributor I

    Hi,

    the logistic regression operator cannot handle any kind of nominal attributes. It can only be applied on data sets with numeric attributes. So you have to use the Nominal to Numerical operator to dummy coding your polynominal attributes.

    Greetings,

    Sigrun

  • phellingerphellinger Employee, MemberPosts:103RM Engineering

    Hi Sigrun,

    the new Logistic Regression algorithm (starting from version 7.2) does handle nominal attributes out of the box.

    Best,

    Peter

Sign InorRegisterto comment.