Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


akron1312341058.pdf (1.35 MB)

ETD Abstract Container

Abstract Header

A Hybrid Classifier Committee Approach for Microarray Sample Classification

Vishnampettai Sridhar, Aadhithya
http://rave.ohiolink.edu/etdc/view?acc_num=akron1312341058

Abstract Details

2011, Master of Science, University of Akron, Computer Science.
Microarray technology is one of the important breakthroughs in the field on bioinformatics. Microarray data is usually considered noisy and has high dimensionality. Thus, performing analysis on microarray data requires fine tuned data mining techniques. Out of many analyses performed on microarray data, classification of biological samples is the most frequent. This study focuses on building a committee of hybrid classifiers to classify microarray samples. For the purpose of study, support vector machine and K-nearest neighbor classifiers were used. Datasets namely, leukemia dataset and cardiac dataset were used for the study. A binary classification was done on the leukemia dataset and three class classification was performed on the cardiac dataset. The three class classification was done in two steps, in each of which, binary classification was employed. For both leukemia and cardiac dataset, a subset of randomly chosen samples was used for validating the committee. The rest of the dataset was used for formation of the committee. For this, the dataset was preprocessed and feature selection was performed. Then it was classified using a group of classifiers and the classifier with highest accuracy was chosen into the committee. The committee thus formed, was then validated using the validation set chosen initially. The results indicate that the committee outperformed at least 3 individual classifiers out of 4. In some cases, it performed better than all the individual classifiers. This proves the consistency of the committee. The confidence level calculated for the committee was about 95% for both leukemia and cardiac datasets, proving its reliability.
Zhong-Hui Duan, Dr (Advisor)
Kathy Liszka, Dr. (Committee Member)
Yingcai Xiao, Dr. (Committee Member)
74 p.
Computer Science
Hybrid Committee Approach

Recommended Citations

Citations

  • Vishnampettai Sridhar, A. (2011). A Hybrid Classifier Committee Approach for Microarray Sample Classification [Master's thesis, University of Akron]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=akron1312341058

    APA Style (7th edition)

  • Vishnampettai Sridhar, Aadhithya. A Hybrid Classifier Committee Approach for Microarray Sample Classification. 2011. University of Akron, Master's thesis. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=akron1312341058.

    MLA Style (8th edition)

  • Vishnampettai Sridhar, Aadhithya. "A Hybrid Classifier Committee Approach for Microarray Sample Classification." Master's thesis, University of Akron, 2011. http://rave.ohiolink.edu/etdc/view?acc_num=akron1312341058

    Chicago Manual of Style (17th edition)

akron1312341058
479
© 2011, all rights reserved.
This open access ETD is published by University of Akron and OhioLINK.