Bug Severity Assessment in Cross Project Context and Identifying Training Candidates

摘要: The automatic bug severity prediction will be useful in prioritising the development efforts, allocating resources and fixer. It needs historical data on which classifiers can trained. In absence of such cross project provides a good solution. this paper, our objective is to automate by using metric summary identify best training candidates context. text mining technique has been used extract terms trained these terms. About 63 have designed combining seven datasets Eclipse projects develop models. To deal with imbalance problem, we employed two approaches ensemble operators available RapidMiner: Vote Bagging. Results show that k-Nearest Neighbour (k-NN) performance better than Support Vector Machine (SVM) performance. Naive Bayes f-measure poor, i.e...