Title :
A Hybrid Feature Selection Method Using Gene Expression Data
Author :
Chuang, Li-Yeh ; Wu, Kuo-Chuan ; Yang, Cheng-Hong
Author_Institution :
Chem. Eng., I-Shou Univ., Kaohsiung, Taiwan
Abstract :
In this paper, correlation-based feature selection (CFS) and the Taguchi-genetic algorithm (TGA) method were combined in a hybrid method, and the K-nearest neighbor (KNN) method with leave-one-out cross-validation (LOOCV) served as a classifier for eleven classification profiles. With the help of this classifier classification accuracy were calculated. Experimental results show that this method effectively simplifies features selection by reducing the total number of features needed. The proposed method obtained the highest classification accuracy in five out of the six gene expression data set test problems when compared to other classification methods from the literature.
Keywords :
Taguchi methods; genetics; molecular biophysics; K-nearest neighbor method; Taguchi-genetic algorithm method; classifier classification accuracy; correlation-based feature selection; eleven classification profiles; gene expression data; hybrid feature selection method; leave-one-out cross-validation; Accuracy; Bioinformatics; Biomedical engineering; Chemical engineering; Computer science; Data engineering; Filters; Gene expression; Genetic algorithms; Testing; Feature selection; K-nearest neighbor; Leave-one-out cross-validation; Taguchi-genetic algorithm;
Conference_Titel :
Bioinformatics and BioEngineering, 2009. BIBE '09. Ninth IEEE International Conference on
Conference_Location :
Taichung
Print_ISBN :
978-0-7695-3656-9
DOI :
10.1109/BIBE.2009.24