Class Based Ensemble Similarity Measure Based Efficient Classification of Medical Data

Abstract

The medical data classification problem has been well studied and there exist numerous approaches which uses different metrics and measures. However, the previous algorithms suffering in classification to achieve expected efficiency. To improve the performance of medical data classification, a CBES (Class Based Ensemble Similarity) based algorithm is presented. The method first computes amount of ensembles for dissimilar group of medical data. Second, for each class with the input sample, a CBES measure has been estimated. The CBES measure has been computed based on the number of complete features match of the ensemble and number of partial feature similarity appeared. Using both of them, the method computes the CBES measure for the given test sample. A single class of diseases has been defined based on the CBES measure. The suggested CBES method has reduced time complexity and enhanced classification performance.