Research On Duplicate Bug Report Prediction Based On Semantic Extended Continuous Query

With the increase in the scale and complexity of software projects,a large number of bug reports have been generated during the testing process,among which duplicate bug reports are common.The existence of duplicate bug reports reduces the efficiency of developers to fix bugs.Duplicate bug report prediction can effectively avoid the generation of duplicate bug reports.It is one of the hot research directions in recent years,but the efficiency and accuracy of duplicate bug report prediction is not high.To this end,a method for predicting duplicate bug reports based on semantic expansion of continuous queries is proposed.By constructing an bug report index vocabulary based on topic models,the query word sequence is expanded based on semantic prediction,and a continuous query duplicate bug report collaborative prediction algorithm is adopted.,While shrinking the index space,it improves the prediction accuracy and efficiency of duplicate bug reports.Experiments show that compared with the traditional duplicate bug report prediction method,this method reduces the bug report index space by more than 50%,improves the prediction effect by 33.6%,and shortens the prediction time by 41%-73%.The main research contents of this paper are as follows:(1)The construction of a bug report index lexicon based on the topic model.By extracting the subject of the bug report,the semantic association between the terms contained in the bug report and the subject of the bug report is constructed,and the subject terms of the bug report are determined based on this,and the bug report index vocabulary based on the topic model is constructed,to a greater extent This greatly reduces the space of the traditional bug report index lexicon.(2)Duplicate bug report prediction method based on semantic expansion continuous query.The duplicate bug report prediction method based on semantic expansion of continuous query mainly includes:query word sequence expansion based on semantic prediction;duplicate bug report collaborative prediction of continuous query;and duplicate bug report prediction result sorting.In the query term sequence expansion based on semantic prediction,the query term sequence after semantic expansion is formed by synonym supplementation of the current query term sequence and prediction of subsequent semantic words.After the query word sequence expansion based on semantic prediction,the recall rate and precision rate of the query can be improved.In the duplicate bug report collaborative prediction of continuous query,the antecedent word of the current input word is used in the bug report,and the expanded sequence of the antecedent query word is obtained after expansion,which reduces the space of prediction search and greatly improves the prediction speed.