School of Computing

Jan 22
14:00 - 14:30
Data Science: Helen Costa Lima
Data Science Group Seminar
Filter and metaheuristic approaches for feature selection in a hierarchical classification context

Abstracts: In this talk, I will explore the use of feature selection techniques for the classification task that take into account the hierarchy of classes. The proposed approaches combine filter techniques (Symmetrical Uncertain and Correlation-based) and metaheuristic techniques (Variable Neighbourhood Search and Genetic Algorithms) to search and evaluate feature subsets to construct solutions capable of improving the predictive performance of global hierarchical classifiers. The development of this type of feature selection will enable researchers to better analyse datasets in a hierarchical classification context, such as the prediction of protein functions.


Intellectual Hub,
Medway Building,
University of Kent,
Chatham Maritime,
United Kingdom


Open to all,

Contact: Daniel Soria
School of Computing

School of Computing, University of Kent, Canterbury, Kent, CT2 7NF

Enquiries: +44 (0)1227 824180 or contact us.

Last Updated: 14/08/2015