Mouse protein classification using tree based machine learning techniques

Abstract

From the last decades, it has been observed that the protein enzyme classification is in the trend and got a lot of attraction from research community. Generally, the protein classification is done by the clinical laboratories with the help of chemical bonds of proteins. Although, this process is very authentic, but it consumes a lot of resources and man power and still a lot of proteins are un-reviewed are not classified in any type. In this context, many computational models have been given by the scholars. This article consider the tree based classification methods viz “CRT, CHAID, C5.0, ANN (Network based) and SVM” used for mouse protein enzyme class categorization and predictions. Total 2282 proteins with 1437 features has been used for experimental analysis and performance evaluation of the model is done by the standard validation techniques viz accuracy, specificity, sensitivity, precision, recall, f-measures and MCC. The computational results said that the SVM achieves 97.08% accuracy and it may be used for categorization and predication.