|
|
|
|
|
|
Article Detail
|
|
Label
|
|
|
Label
|
|
|
|
AN EFFECTUAL FILTER BASED GENE SELECTION WITH DNCM-IPSO ALGORITHM FOR DIAGNOSIS OF PERIPHERAL BLOOD CELLS(PBCs) IN RHEUMATOID ARTHRITIS (RAs)
|
|
| Author:
|
B. 2, DR.R. NEDUNCHEZHIAN
|
| Abstract:
|
Rheumatoid Arthritis (RA) is an illness of chronic inflammatory arthritis. Presently, diagnosing RA may involve number
of weeks, and the factor applied for the prediction of a poor prognosis is not trustable at all time. Gene expression in
RA may contain a distinct signature. Gene expression analysis has been used for synovial tissue for defining
molecularly unique forms of RA; but, the expression analysis of tissue obtained from a synovial joint is intrusive and
clinically not feasible. The research carried out recently have shown that distinct gene expression variations can be
found in Peripheral Blood Mononuclear Cells (PBMCs) taken from different patients affected with cancer, multiple
sclerosis and lupus. In order to find RA disease-associated genes, a gene selection and classification was performed. At
first, with the result of reducing the time complexity, this dataset related to the illness is procured and then the data
is preprocessed. Subsequently, to reduce the number of genes, the gene data is selected from the preprocessed data
with the help of filter dependent gene selection methods, which include: T-test, chi-squared test, relief-F and
Minimum Redundancy Maximum Relevance (mRMR). Thirdly, Enhance Entropy with Gaussian Kernel based Support
Vector Machine (EEGK-SVM) approach is proposed for disease prediction, in turn, maximizes the prediction accuracy.
At last, for the RA disease classification, a Dynamic Neutrosophic Cognitive Map with Improved Particle Swarm
Optimization (DNCM-IPSO) algorithm is introduced, which is quite suitable for the medical routine and it is
presented for aiding the gene expression in the early and accurate diagnosis of RA patients. Consequently, RA
disease is not allowed from getting into advanced stages and the hardship with developing insistent and also erosive
arthritis for RA patients will also be decreased. The results prove that the DNCM-IPSO methodology performs
better performance in terms of precision, accuracy, recall and F-measure etc when compared with the other
algorithms.
|
| Keyword:
|
Rheumatoid Arthritis (RAs), gene expression profiling, Gene selection, filter based gene selection, Enhance Entropy with Gaussian Kernel, Support Vector Machine (EEGK-SVM), Dynamic Neutrosophic Cognitive Map (DNCM), and Improved Particle Swarm Optimization (IPSO).
|
| EOI:
|
-
|
| DOI:
|
https://doi.org/10.31838/ijpr/2018.10.04.019
|
| Download:
|
Request For Article
|
|
|
|
|
|
|
|
|
|
|