An effective rank approach to software defect prediction using software metrics

Abstract

An important goal throughout the cycle of software system development is to search out and fix existing defects as early as attainable. This has abundant to do with software system defects prediction and management. There are square measure primarily two classes among these prediction models. One class is to predict what percentage defects still exist in keeping with the already captured defects knowledge within the earlier stage of the software system's life-cycle. The opposite class is to predict what percentage defects will be there within the newer version software system in keeping with the sooner version of the software system defects knowledge. Within the present situation, defect prediction is predicated solely on the dimensions that's supported LOC count that is not abundant economical. This paper is all concerning predicting defects with the exploitation of object oriented metrics and version history for every module. Once the prediction method is over, the modules square measure being stratified in keeping with their severity and therefore the overall price for the trouble is calculable.