Code Smell Aware – Issue Prediction Model

S.Narasimhulu, CH.Lawrence Dheeraj, Dr.Madhu B.K

Issues are the reasons for poor design. Previously we assess the impact of smells on code quality and it indicates their harmful impact on maintainability. In this paper we collect previous detections on issue-proneness to construct a specialized issue prediction model for code smell classes. Mainly focus on the involvement of a measure the severity of the code smells by adding it to the existing issue prediction model product based process based metrics, and comparing the results of the new model. The proposed model with the one of alternative approach which impacts metrics about the previous data of code smells in files. Identify that one proposed works usually better. However we observed the complementarities between the set of issues and smelly classes properly classified by the two models. On the basis of this result we assess a smell aware combined issue prediction model. We make obvious how such model classifies issue-prone code components with the harmonic mean of precision and recall.