Citation

Goh, Pey Yun and Tan, Shing Chiang (2020) Identifying and Removing Outlier Features Using Neighborhood Rough Set. In: Information Science and Applications. Lecture Notes in Electrical Engineering (Information Science and Applications), 621 . Springer, pp. 485-495. ISBN 9789811514647

Text 10.1007@978-981-15-1465-4.pdf - Published Version Restricted to Repository staff only Download (67MB)

Abstract

The neighborhood rough set (NRS) is used to remove redundant features after identifying neighborhood relation among samples of features. In this study, a new NRS is proposed to determine and remove outlier features. An outlier score is calculated by measuring the neighborhood relation and non-neighborhood relation among samples with respect to a feature. Features that have an outlier score below the average outlier score are removed from the data set. In this research work, a support vector machine (SVM) and its extended version to reduce input features are used to evaluate the quality of the selected features from the proposed NRS. The experiment involves twelve real world data sets. The results show that the proposed method can reduce at least half of the features effectively from these data sets. Although the classification accuracy is slightly lower than both SVM-based solutions, the proposed NRS with SVM could significantly remove more number of input attributes and requires much shorter execution time.

Item Type:	Book Section
Uncontrolled Keywords:	Outlier, Neighborhood rough set, Features selection
Subjects:	Q Science > QA Mathematics > QA273-280 Probabilities. Mathematical statistics
Divisions:	Faculty of Information Science and Technology (FIST)
Depositing User:	Ms Suzilawati Abu Samah
Date Deposited:	14 Dec 2020 16:37
Last Modified:	14 Dec 2020 16:37
URII:	http://shdl.mmu.edu.my/id/eprint/7938

Downloads

Edit (login required)