Applying bug features based clustering for prediction of bug fixing time

Kwanghue Jin, Kimun Kwon, Byungjeong Lee

Research output: Contribution to journalArticlepeer-review


Fixing bugs is the one of the important task in the field of software maintenance. Prediction of bug fixing time helps software engineers plan time and cost in the process of software development and maintenance. It has also the positive impact on software quality. Therefore, in this paper we propose a technique to predict the bug fixing time for planning software projects and improving software quality. In order to predict the bug fixing time, we utilize seven features of bug reports and group bug reports. KNN method is used for grouping. And we compute text similarity between bug reports to predict fixing time with more accuracy. Comment and description field of bug reports are used for the computation of text similarity. We introduce several measures for verification of prediction accuracy such as Average Absolute Residual (AAR), Pred(x), and Magnitude Relative Error (MRE). In experiments we apply these measures to Open Source Projects and evaluate our technique.

Original languageEnglish
Pages (from-to)5577-5584
Number of pages8
Issue number11A
StatePublished - 1 Nov 2014


  • Bug fixing
  • Bug reports
  • Software maintenance
  • Time prediction


Dive into the research topics of 'Applying bug features based clustering for prediction of bug fixing time'. Together they form a unique fingerprint.

Cite this