What makes two files change together? Can we predict when they are going to co-change again?

We investigated this question on the Derby project. We collected data from issue reports, discussions, and commits and used machine learning techniques to build a prediction model. We analyzed 5266 commits in which a pair of files changed together and predicted correctly 86% of them.

The most useful information considered by the prediction model was:

• number of lines of code added,
• number of lines of code removed,
• sum of number of lines of code added, modified and removed,
• number of words used to describe and discuss the issues, and
• median value of closeness, a social network measure obtained from issue comments.

The table below summarize the co-change prediction results by version. Please, inspect the results for files that might be more interesting for you.
The columns are:

• # of Co-changes: is the number of commits collected when file 1 changed together with file 2,
• # of No Co-changes: is the number of commits collected when file 1 changed without file 2,
• File 1: is the file firstly committed. We used metrics collected from this first commit to predict the co-change with file 2,
• File 2: is the file that was committed after the File 1,
• Correct Co-Changes: is the number of commits that we got correct predictions when file 1 changed with File 2,
• False Alarm: is the number of commits that we predicted that File 2 would change together with File 1, but it didn’t,
• Wrong: is the number of commits that we did not predict that both files would co-change,
• Correct No Co-Changes: is the number of commits that we got correct predictions when file 1 changed without File 2.

Do these results surprise you? Can you think in any explanation for the results?
Do you think that our rate of prediction is good enough to be used for building tool support for the software community?
Do you have any suggestion on what can be done to improve the change recommendation?

Feel free to enter in contact with us: igor.wiese@gmail.com.
Our research group is very interested in helping developers to make their software changes.
All the best,
Igor Wiese
Phd Candidate