Seminar on Computational Learning and Adaptation

 
Predictive Clustering Trees and Their Use in Bioinformatics

Saso Dzeroski
Department of Knowledge Technologies
Jozef Stefan Institute
Ljubljana, Slovenia
        http://www-ai.ijs.si/SasoDzeroski/SasoDzeroski.html


Predictive clustering trees are an approach to data analysis that include elements of both clustering and predictive modelling. They allow for conceptual clustering, where we seek both clusters and their descriptions, as well as simultaneous prediction of several target variables.

The talk will introduce predictive clustering trees, present methods for learning them and summarize some of their applications in Bioinformatics. The areas of Bioinformatics were we have obtained results includes predicting gene functions and relating gene expression profiles to the  clinical status of patients.

This talk is co-sponsored with the Stanford Medical Informatics (SMI) Colloquium Series

Date: Thurs, May 25

Time: 12:15-1:15PM 

Place: SMI Conference Room, MSOB X-275

Return to the seminar schedule