|
|
Bimonthly Since 1986 |
ISSN 1004-9037
|
|
|
|
|
Publication Details |
Edited by: Editorial Board of Journal of Data Acquisition and Processing
P.O. Box 2704, Beijing 100190, P.R. China
Sponsored by: Institute of Computing Technology, CAS & China Computer Federation
Undertaken by: Institute of Computing Technology, CAS
Published by: SCIENCE PRESS, BEIJING, CHINA
Distributed by:
China: All Local Post Offices
|
|
|
|
|
|
|
|
|
|
Abstract
In the field of medical science, feature (or gene) selection is a burning topic. Microarray data are very important for the feature selection to diagnose any disease. It may be possible that few thousands of features are present in microarray data. That’s the reason why to find out a proper subset of features using conventional algorithms, it becomes an important object. We have to reduce the whole dataset’s dimension in order to produce a proper subset, keeping in mind that we don’t have to miss significant features while removing redundant features. For disease diagnosis small features set with minimal features. Autoencoder technique is proved to be very powerful and efficient for reducing the dimension. Inspired by the Autoencoder technique, both of us built a model based on Folded Autoencoder (FA) for the selection of features set. After that, some deep learning classifiers are applied to check the accuracy of classifier. Performance of Support Vector Machine (SVM) is better than other classifiers after reducing the dimension of features. This model is named as Folded Autoencoder – SVM (FAS). Lastly, there is a comparison of results obtained from whole dataset (without applying FA) and reduced dataset (after applying FA).
Keyword
PR v1.0, Microarray Data Analysis, Genes, Feature Selection, Folded Auto-encoder, SVM, KNN
PDF Download (click here)
|
|
|
|
|