|Title:||On the data imbalance problem in SVM-based subcellular localization|
|Subject:||Support vector machines|
Hong Kong Polytechnic University -- Dissertations
|Department:||Department of Electronic and Information Engineering|
|Pages:||37 leaves : col. ill. ; 30 cm.|
|Abstract:||Several computational methods have been proposed for protein subcellular localization. SVMs perform well when classes are equally represented, but are limited in their performance on highly imbalanced datasets. Here we propose two novel methods to overcome this data-imbalance problem. The proposed methods use over-sampling techniques to create new minority-class patterns to rebalance the datasets. Pairwise alignment is used to align query sequences with training sequences and obtain score vectors with dimensionality equal to the number of training sequences. Experimental results on two subcellular localization datasets show that the proposed over-sampling methods are able to improve the total accuracy. We believe that the proposed methods will not only be applicable to subcellular localization, but also to other machine learning applications.|
|Rights:||All rights reserved|
Files in This Item:
|b24468381.pdf||For All Users (off-campus access for PolyU Staff & Students only)||955.49 kB||Adobe PDF||View/Open|
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item: