DI NUOVO, Alessandro, PALESI, Maurizio and CATANIA, Vincenzo (2007). Multi-objective evolutionary fuzzy clustering for high-dimensional problems. In: 2007 IEEE International Fuzzy Systems Conference , 23-26 July 2007. IEEE, 1-6.
1495_camera_ready.pdf - Accepted Version
Available under License All rights reserved.
Download (155kB) | Preview
This paper deals with the application of unsupervised fuzzy clustering to high dimensional data. Two problems are addressed: groups (clusters) number discovery and feature selection without performance losses. In particular we analyze the potential of a genetic fuzzy system, that is the integration of a multi-objective evolutionary algorithm with a fuzzy clustering algorithm. The main characteristic of the integrated approach is the ability to handle the two problems at the same time, suggesting a Pareto set of trade-off solutions which could have a better chance of matching the real needs. We exhibit the high quality clustering and features selection results by applying our approach to a real-world data set.
|Item Type:||Book Section|
|Additional Information:||Poster originally presented at IEEE International Fuzzy Systems Conference, 2007. FUZZ-IEEE 2007. 23-26 July 2007, London, UK. ISSN : 1098-7584|
|Research Institute, Centre or Group:||Cultural Communication and Computing Research Institute > Communication and Computing Research Centre|
|Depositing User:||Alessandro Di Nuovo|
|Date Deposited:||26 Jul 2016 13:36|
|Last Modified:||19 Oct 2016 17:29|
Actions (login required)
Downloads per month over past year