Skip to Main Content
High dimensionality of microarray datasets pose a great challenge to the researchers classifying them. Traditional classifiers perform poorly on these datasets because of their large, redundant and irrelevant feature set. Therefore a small number of features (genes) are always desirable both by the classifier to classify better as well as to the biologist to analyze the cause of disease with fewer important genes. In this study we have proposed an efficient feature selection technique based on linear regression analysis which finds the best feature set using a regression model. The acceptance of the method is evaluated by comparing with several other feature selection approaches in different classifiers.