Jerry Friedman's RuleFit finds "important rules", which I guess can be thought of as features...
Andy From: Weiwei Shi > > "The goal of logic regression is to find predictors that are Boolean > (logical) combinations of the original predictors", which > might be more like what I need. > > The rf's variable importance evaluations and oob might help > feature selection but not feature construction, which is more > of my concern. > > Thanks for any further suggestion or explanation. > > On 4/7/06, Sean Davis <[EMAIL PROTECTED]> wrote: > > > > Weiwei, > > > > You could also look into using one of several methods of > > "classification" that calculate the "weight" of individual > predictors > > in producing a correct result based on some version of > > cross-validation. One that I use relatively > > often is randomForest (in the randomForest package). > > > > Sean > > > > > > > > On 4/7/06 12:34 PM, "Peter Ehlers" <[EMAIL PROTECTED]> wrote: > > > > > It sounds as though 'logic regression' might help. See > > > > > > Ruczinski I, Kooperberg C, LeBlanc ML (2003). Logic Regression, > > > Journal of Computational and Graphical Statistics, 12, 475-511. > > > > > > and the LogicReg package. > > > > > > Peter Ehlers > > > > > > Weiwei Shi wrote: > > > > > >> Hi there, > > >> I have a statistics question on a classification problem: > > >> > > >> Suppose I have 1000 binary variables and one binary dependent > > >> variable. > > I > > >> want to find a way similar to PCA, in which I can find a > couple of > > >> combinations of those variables to discriminate best > according to > > >> the dependent variable. It is not only for dimension > reduction, but > > >> more important, for finding best way to construct > features. This is > > >> NOT CDA > > since > > >> the explanatory variables are NOT continuous. I knew the > existence > > >> of > > that > > >> method since I consulted before with a professor but I > forgot the > > >> name > > of > > >> the method.. sigh... > > >> > > >> I am also wondering if R has already some function or package > > addressing > > >> this kind of problem. > > >> > > >> Thanks > > >> > > >> -- > > >> Weiwei Shi, Ph.D > > >> > > >> "Did you always know?" > > >> "No, I did not. But I believed..." > > >> ---Matrix III > > >> > > >> [[alternative HTML version deleted]] > > >> > > >> ______________________________________________ > > >> [email protected] mailing list > > >> https://stat.ethz.ch/mailman/listinfo/r-help > > >> PLEASE do read the posting guide! > > http://www.R-project.org/posting-guide.html > > > > > > ______________________________________________ > > > [email protected] mailing list > > > https://stat.ethz.ch/mailman/listinfo/r-help > > > PLEASE do read the posting guide! > > http://www.R-project.org/posting-guide.html > > > > > > > -- > Weiwei Shi, Ph.D > > "Did you always know?" > "No, I did not. But I believed..." > ---Matrix III > > [[alternative HTML version deleted]] > > ______________________________________________ > [email protected] mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide! > http://www.R-project.org/posting-guide.html > > ______________________________________________ [email protected] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
