Loading...

Gene Selection and Reduction in DNA Microarrays to Improve Classification Accuracy of Cancerous Samples

Baharvand Irannia, Zohreh | 2010

437 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 40119 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Rabiee, Hamid Reza
  7. Abstract:
  8. DNA Microarray is the state-of-the-art technology in analyzing gene expression data. It has made it possible to measure expression levels of thousand of genes simultaneously. Microarray classification has been widely used in effective diagnosis of cancers and some other biological diseases. But the most challenging issue is the intense asymmetry between the dimensionality of genes and tissue samples which can wreck the classification performance. This dissertation will focus on gene selection and reduction solutions and presents a novel classification scheme which uses both gene selection and dimension reduction in its different stages. We have improved one of the recently proposed topology preserving reduction methods, locality preserving projections (LPP), and have used it in its both unsupervised and supervised forms. We constructed the unsupervised LPP’s affinity matrix in a new way that improves the quality of the selected genes and the classification performance in further steps. The proposed method has been tested on six publicly available cancer microarray datasets. Experimental results show that the suggested method greatly outperforms the similar competing methods which use simpler reduction approaches or naïve forms of LPP. The dissertation concludes that the proposed novel topology preserving method can highly eliminate irrelevant and redundant genes and classify microarray data in the reduced space effectively.

  9. Keywords:
  10. Machine Learning ; DNA Microarrays ; Dimension Reduction ; Cancer ; Bioinformatics ; Gene Expression Data ; Classification ; Accuracy ; Gene Selection

 Digital Object List

  • محتواي پايان نامه
  •   view

 Bookmark

No TOC