Subgroup discovery and rule evaluation in ROC space
Report a problem or upload filesIf you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.
Traditionally, machine learning has focussed on induction of classification and prediction rules. More recently, non-predictive or descriptive induction is gaining substantial interest of machine learning researchers. Two major trends in descriptive induction are association rule learning and subgroup discovery. In this seminar we present our recent work in descriptive induction.
We also argue that accuracy is not always an appropriate evaluation measure in the descriptive induction framework, and propose quality measures designed for subgroup evaluation in ROC space. After a brief presentation of the APRIORI-C and SD-algorithm, we give a detailed presentation of the CN2-SD algorithm, which includes a new -- weighted -- covering algorithm, a new search heuristic (weighted relative accuracy), probabilistic classification of instances, and a new measure for evaluating the results of subgroup discovery (area under ROC curve).
The presented work was done in collaboration with V. Jovanoski (APRIORI-C), D. Gamberger (SD-algorithm), B. Kavsek and L. Todorovski (CN2-SD algorithm).
Our research was supported by the Slovenian Ministry of Education, Science and Sport, the IST-1999-11495 project Data Mining and Decision Support for Business Competitiveness: A European Virtual Enterprise, and the British Council project Partnership in Science PSP-18.
Link this pageWould you like to put a link to this lecture on your homepage?
Go ahead! Copy the HTML snippet !