I am familiar with the rpart and tree packages for classification and regression trees. However, quite a bit of the research in the transportation community relating to decision trees uses the C4.5 family of algorithms by Quinlan. Are there any plans to make a C4.5 (or a derivative of it) available to R? If not, then I might use the WEKA Java package ( http://www.cs.waikato.ac.nz/ml/weka) that implements a Java version of C4.5. Thanks. Benjamin Stabler Transportation Planning Analysis Unit Oregon Department of Transportation 555 13th Street NE, Suite 2 Salem, OR 97301 Ph: 503-986-4104
Please check the licence conditions for C4.5: it is basically commercial. You are not allowed to use the published source code, and I am not allowed to show you the source code I bought, which I am not allowed to `use for commercial purposes or gain'. WEKA includes a re-implementation of the ideas behind C4.5, but not C4.5. On Mon, 8 Mar 2004 Benjamin.STABLER at odot.state.or.us wrote:> I am familiar with the rpart and tree packages for classification and > regression trees. However, quite a bit of the research in the > transportation community relating to decision trees uses the C4.5 family of > algorithms by Quinlan. Are there any plans to make a C4.5 (or a derivative > of it) available to R? If not, then I might use the WEKA Java package ( > http://www.cs.waikato.ac.nz/ml/weka) that implements a Java version of C4.5. > Thanks.-- Brian D. Ripley, ripley at stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UK Fax: +44 1865 272595
Hi,> -----Original Message----- > > WEKA includes a re-implementation of the ideas behind C4.5, > but not C4.5.If my memory serves me right, WEKA people called this "re-implementation" J4.8. Kevin -------------------------------------------- Ko-Kang Kevin Wang, MSc(Hon) Statistics Workshops Co-ordinator Student Learning Centre University of Auckland New Zealand
Thanks for the insight. I think we will try both CART (rpart) and C4.5 (J4.8) and see what happens. As always, the R community is so helpful. Thanks, Ben>-----Original Message----- >From: Ko-Kang Kevin Wang [mailto:k.wang at auckland.ac.nz] >Sent: Tuesday, March 09, 2004 12:08 AM >To: 'Prof Brian Ripley'; STABLER Benjamin >Cc: r-help at stat.math.ethz.ch >Subject: RE: [R] Decision Trees > > >Hi, > >> -----Original Message----- >> >> WEKA includes a re-implementation of the ideas behind C4.5, >> but not C4.5. > >If my memory serves me right, WEKA people called this >"re-implementation" J4.8. > >Kevin > >-------------------------------------------- >Ko-Kang Kevin Wang, MSc(Hon) >Statistics Workshops Co-ordinator >Student Learning Centre >University of Auckland >New Zealand > >