An efficient filter based classification approach for microarray disease detection

As the size of the biomedical databases areincreasing day-by-day, finding anessential featureset for classification problem is complex due to large data size and sparsity problems. Microarray feature ranking and classification is one of the major challenges to scientific and medical researchers due to its high dimensional feature space and limited number of samples. Feature transformation, feature ranking and data classification are the essential components to improve the microarray cancer prediction on high dimensional datasets. In this work, a novel framework is designed and implemented to classify the high dimensional data with high true positive rate. In the proposed work, a hybrid feature transformation, hybrid feature selection and advance classification approach are implemented to improve the true positive rate and error rate of the disease prediction. A novel principal component ranking measure is integratedin order to find the subset of features for classification problem. Finally, a hybrid decision tree classifier is used to predict the classification accuracy on the selected features set. Experimental results proved that the present framework has better performance compared to the traditional models for variable microarray datasets