A match count method (MCM) for feature selection with cancer datasets in a neuro-fuzzy system
- Authors
- Lim, J.; Shin, B.; Lim, J.S.
- Issue Date
- Aug-2017
- Publisher
- Institute of Advanced Scientific Research, Inc.
- Keywords
- Biomarker; Feature Selection; Microarray Data; Neuro Network Fuzzy Algorithm
- Citation
- Journal of Advanced Research in Dynamical and Control Systems, v.9, no.8, pp.121 - 127
- Journal Title
- Journal of Advanced Research in Dynamical and Control Systems
- Volume
- 9
- Number
- 8
- Start Page
- 121
- End Page
- 127
- URI
- https://scholarworks.bwise.kr/gachon/handle/2020.sw.gachon/80287
- ISSN
- 1943-023X
- Abstract
- Background/Objectives: Finding the appropriate features for the specific target such as a colon cancer has been a challenging issue in bioinformatics. Methods/Statistical analysis: We propose a Match-Count Method (MCM) for selecting appropriate features that can be candidate biomarkers from the colon cancer microarray dataset and minimizing the number of biomarkers from the extracted candidate features with the neural network with weighted fuzzy membership function(NEWFM).We compared our proposed method with Gini Index, Chi Square and Maximum Relevance-Minimum Redundancy (MRMR) in terms of the accuracy with the selected features. Findings: We use the colon dataset from the public Kent Ridge Bio-medical Data Repository for our comparative experiments. The accuracy of proposed method (MCM) was compared with three classifiers. The first classifier is a Bayesian classifier that is a representative method as a statistical measure. The second classifier is J48 classifier that is Weka’s implementation of C4.5 algorithm that is the induction of decision trees. The last classifier is a neural network classifier that is our previous proposed classifier. The proposed method showed the highest accuracy compared to Gini Index, Chi Square and MRMR. Finally, we selected the minimum number of features (attribute1560, 767, 377, 1924) with the highest accuracy (95.16%) with this given colon dataset. In terms of accuracy, our proposed method showed the highest accuracy. The comparative experimental results showed that the proposed method selected the minimum number of appropriate features that can be the biomarkers in the colon cancer datasets. And those results also improved higher accuracy of selecting appropriate features. Improvements/Applications: The comparative experimental results showed that the proposed method selected the minimum number of appropriate features that can be the biomarkers in the colon cancer dataset and improved higher accuracy of selecting appropriate features. © 2017, Institute of Advanced Scientific Research, Inc.. All rights reserved.
- Files in This Item
- There are no files associated with this item.
- Appears in
Collections - IT융합대학 > 컴퓨터공학과 > 1. Journal Articles
![qrcode](https://api.qrserver.com/v1/create-qr-code/?size=55x55&data=https://scholarworks.bwise.kr/gachon/handle/2020.sw.gachon/80287)
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.