Project: A Matrix Factorization approach to Exceptional Model Mining
Description
Exceptional Model Mining aims to identify subgroups in the dataset that behave somehow exceptionally. It differs from a clustering approach since subgroups may overlap; not all data points are assigned to a cluster. However, consequently, the list of subgroups often contains many similar, redundant subgroups, which is not meaningful for domain experts.
In this project, we aim to develop an optimization technique to find one, optimal set of subgroups. We expect to need matrix factorization, gradient descent and other optimization techniques. We build on an existing idea that can be the starting point, and which needs to be further developed and experimentally evaluated.
For more information and to see if this project is a match for you, please do not hesitate to reach out: r.m.schouten@tue.nl.
Details
- Supervisor
-
Rianne Schouten
- Secondary supervisor
-
Sibylle Hess
- Interested?
- Get in contact