Title :
A Filter Correlation Method for Feature Selection
Author :
Hosni, Hanen ; Mhamdi, Faouzi
Author_Institution :
Nat. Super. Sch. of Eng. of Tunis, Univ. of Tunis, Tunis, Tunisia
Abstract :
Biological data is undergoing exponential growth in volume and complexity. Often, the selection of biological features is a crucial step that aims to defy the curse of dimensionality to improve prediction performance in classification systems, facilitate viewing, understanding and analyzing data. In this paper we present an adaptation of the Fast Correlation Based Filter algorithm (FCBF) whose aims is to identify relevant, not redundant features to improve the capacity of prediction and reduce the search space.
Keywords :
bioinformatics; data mining; feature selection; information filtering; FCBF; biological data; biological feature selection; classification system prediction performance improvement; curse of dimensionality; data analysis; data understanding; data viewing; fast correlation based filter algorithm; filter correlation method; prediction capacity improvement; search space reduction; Algorithm design and analysis; Biology; Classification algorithms; Correlation; Erbium; Filtering algorithms; Support vector machines; KDD; bioinformatics; biological macromolecules; correlation; feature selection; filter approach;
Conference_Titel :
Database and Expert Systems Applications (DEXA), 2014 25th International Workshop on
Conference_Location :
Munich
Print_ISBN :
978-1-4799-5721-7
DOI :
10.1109/DEXA.2014.28
