Re: [Scikit-learn-general] Using subgroup discovery

2014-09-01 Thread Debanjan Bhattacharyya
Hi Gael I would say the Seminal Paper is the following Wrobel, S. (1997). An algorithm for multi-relational discovery of subgroups. In Principles of Data Mining and Knowledge Discovery, pages 78{87. Springer. And more papers are here Abudawood, T. and Flach, P. (2009). Evaluation measures for m

Re: [Scikit-learn-general] Using subgroup discovery

2014-09-01 Thread Gael Varoquaux
On Mon, Sep 01, 2014 at 11:38:46PM +0530, Debanjan Bhattacharyya wrote: > If you type "Cortana Subgroup Discovery" in scholar.google.com, you will get a > list of papers. Which is the seminal paper. I get only 15 papers with this query, and none have much citations. Cheers, Gaël ---

Re: [Scikit-learn-general] Using subgroup discovery

2014-09-01 Thread Debanjan Bhattacharyya
Hi Gaël Subgroup Discovery is a well established algorithm. If you type "Cortana Subgroup Discovery" in scholar.google.com, you will get a list of papers. Some more can be seen here http://datamining.liacs.nl/background.html I think it will satisfy all the conditions mentioned in the link you p

Re: [Scikit-learn-general] Using subgroup discovery

2014-09-01 Thread Gael Varoquaux
On Mon, Sep 01, 2014 at 11:01:25PM +0530, Debanjan Bhattacharyya wrote: > Subgroup Discovery is a great option to bridge the gap here (like we have > Cortana implemented in Java. Python will be much faster without a doubt.). > Is there any plan to get this incorporated within sklearn ? The necess

[Scikit-learn-general] Using subgroup discovery

2014-09-01 Thread Debanjan Bhattacharyya
Hi All I have been working for a while on sklearn, targeting different real life machine learning problems. I have experienced there is a large range of problems were the tree algorithms are not at par, because of the way the splits are done, (gt and lt) on values. This is specifically applicable