|
|
 |
| |
|
|
|
A Novel Feature Selection Algorithm for Coronary Artery Disease Prediction |
|
|
|
PP: 735-744 |
|
|
|
doi:10.18576/amis/120408
|
|
|
|
Author(s) |
|
|
|
K. Uma Maheswari,
A. Valarmathi,
|
|
|
|
Abstract |
|
|
| Cardiovascular disease is the predominant cause of death throughout the world. In the present era, many diseases are
caused by gene transformation. Asian Indians have a higher chance of having cardiovascular disease as compared to any other global
population. Identifying relevant and candidate genes for classification of samples is a tedious task when dealing with gene expression
data analysis. The objective of this paper is to find the relevant genes responsible for causing coronary artery disease. In this paper
we developed a novel feature selection algorithm based on fold change and p-value. Instead of selecting genes randomly our proposed
method selects the top ranking candidate genes responsible for coronary artery disease. The selected differentially expressed genes
from the feature selection phase are evaluated using the proposed ensemble classifier. The classifier used in this work are support vector
machine, neural network and naĻıve bayes. The proposed framework is validated by experiments on three publicly available microarray
datasets. The results clearly show that the proposed ensemble classifier performs better when compared to other classifiers. The selected
candidate genes are used for carrying out diagnostic tests and for classifying the patients, which reduces the cost and also improves the
accuracy. |
|
|
|
|
|
|
|
