Brunel University Research Archive (BURA) >
Schools >
School of Engineering and Design >
School of Engineering and Design Research papers >

Please use this identifier to cite or link to this item: http://bura.brunel.ac.uk/handle/2438/1129

Title: An incremental approach to MSE-based feature selection
Authors: Guan, SU
Qi, Y
Bao, C
Keywords: Feature selection
Classifier
Neural network
Feedforward neural network
Minimum squared error (MSE)
Incremental training
Input attribute
Publication Date: 2007
Publisher: World Scientific
Citation: International Journal of Computational Intelligence and Applications. In press
Abstract: Feature selection plays an important role in classification systems. Using classifier error rate as the evaluation function, feature selection is integrated with incremental training. A neural network classifier is implemented with an incremental training approach to detect and discard irrelevant features. By learning attributes one after another, our classifier can find directly the attributes that make no contribution to classification. These attributes are marked and considered for removal. Incorporated with a Minimum Squared Error (MSE) based feature ranking scheme, four batch removal methods based on classifier error rate have been developed to discard irrelevant features. These feature selection methods reduce the computational complexity involved in searching among a large number of possible solutions significantly. Experimental results show that our feature selection methods work well on several benchmark problems compared with other feature selection methods. The selected subsets are further validated by a Constructive Backpropagation (CBP) classifier, which confirms increased classification accuracy and reduced training cost.
URI: http://bura.brunel.ac.uk/handle/2438/1129
ISSN: 1469-0268
Appears in Collections:School of Engineering and Design Research papers
Electronic and Computer Engineering

Files in This Item:

File Description SizeFormat
An Incremental Approach to MSE based Feature Selection.pdf194.69 kBAdobe PDFView/Open

Items in BURA are protected by copyright, with all rights reserved, unless otherwise indicated.

 


Library (c) Brunel University.    Powered By: DSpace
Send us your
Feedback. Last Updated: September 14, 2010.
Managed by:
Hassan Bhuiyan