Gene selection for classifying patients using fuzzy importance factor

Gene selection from microarray data is a widely used technique for obtaining candidate genes in various cancer studies. An important aspect in evaluating a selection method is the similarity among the genes belong to the same class. Experts instinctively have high confidence in the result of a selection method that selects similar sets of genes under some variations to the samples. However, a common problem of existing feature selection methods from microarray gene expression is that the selected genes often vary significantly with sample variations. The paper presents a feature selection method to classify disease and non disease patients based on the similarity in the genes. Similarity is measured using fuzzy importance factor that identify the genes due to which abnormality has been observed in the patients. The identified genes having high multiple interdependence with other genes within the same class are significant for classification of patients. To evaluate the performance of the proposed approach, two well-known microarray data set are considered and the results are compared with other feature selection methods. High classification rate has been achieved with the selected genes obtained using the proposed method.