Predicng	  Reassignments	  of	  Bug	  Reports	  —	  an	  Exploratory	  InvesgaonAhmed Lamkanfi, Serge DemeyerAnsymoAntwerp ...
DevelopersEnd-usersWednesday 12 June 2013
Report through the bug-tracking systemDevelopersEnd-usersWednesday 12 June 2013
Limited technical backgroundReport through the bug-tracking systemDevelopersEnd-usersWednesday 12 June 2013
Provides vital informationLimited technical backgroundReport through the bug-tracking systemDevelopersEnd-usersWednesday 1...
Provides vital informationLimited technical backgroundReport through the bug-tracking systemDevelopersEnd-usersWednesday 1...
Provides vital informationLimited technical backgroundReport through the bug-tracking systemDevelopersEnd-usersPotential i...
Provides vital informationLimited technical backgroundReport through the bug-tracking systemDevelopersEnd-usersComponent a...
/10Product JDTComponentCoreSeverity NormalSummary“Dangerous Eclipsebehavior on file savewhen out of disk space”3Wednesday 1...
/10Product JDTComponentCoreSeverity NormalSummary“Dangerous Eclipsebehavior on file savewhen out of disk space”Product Plat...
/10Product No. of Bugs No. of ComponentsPlatform 35.378 21PDE 7.136 5JDT 16.472 6CDT 7.750 19GEF 852 7Core 54.999 116Bugzi...
/10Product No. of Bugs No. of ComponentsPlatform 35.378 21PDE 7.136 5JDT 16.472 6CDT 7.750 19GEF 852 7Core 54.999 116Bugzi...
/10Cases	  under	  Study:	  Eclipse	  and	  MozillaProduct No. of Reassigned % of ReassignedPlatform 8326 23.5PDE 1290 18....
/10Cases	  under	  Study:	  Eclipse	  and	  MozillaProduct No. of Reassigned % of ReassignedPlatform 8326 23.5PDE 1290 18....
/101. Many reports are assigned to wrong component6Wednesday 12 June 2013
/101. Many reports are assigned to wrong component2. More bug triaging steps needed6Wednesday 12 June 2013
/101. Many reports are assigned to wrong component2. More bug triaging steps needed3. Decrease of productivity6Wednesday 1...
/101. Many reports are assigned to wrong component2. More bug triaging steps needed3. Decrease of productivity✓ Tool suppo...
/10History of all reported bugsUncover/mine facts about historyMake predictions about futureBug Database7Wednesday 12 June...
/10Can we predict which bugs will be assigned to adifferent component?8Wednesday 12 June 2013
/10Can we predict which bugs will be assigned to adifferent component?Binary classification problem8Wednesday 12 June 2013
/10Can we predict which bugs will be assigned to adifferent component?Binary classification problemSummary and initial comp...
/10Can we predict which bugs will be assigned to adifferent component?Binary classification problemSummary and initial comp...
/10Can we predict which bugs will be assigned to adifferent component?Binary classification problemSummary and initial comp...
/10Can we predict which bugs will be assigned to adifferent component?Binary classification problemEvaluation using 10-fold...
/10Accuracy	  ResultsReassignedReassigned No ReassignmentNo ReassignmentProduct Precision Recall Precision RecallPlatform ...
/10Accuracy	  ResultsReassignedReassigned No ReassignmentNo ReassignmentProduct Precision Recall Precision RecallPlatform ...
/10Conclusions10Wednesday 12 June 2013
/10Conclusions10Bugs are often assigned to the wrongcomponentWednesday 12 June 2013
/10Conclusions10Bugs are often assigned to the wrongcomponentOpportunity to improve bug triaging processby predicting comp...
Upcoming SlideShare
Loading in...5
×

Predicting Reassignments of Bug Reports — an Exploratory Investigation

221

Published on

Predicting Reassignments of Bug Reports — an Exploratory Investigation

Published in: Education, Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
221
On Slideshare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
0
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Predicting Reassignments of Bug Reports — an Exploratory Investigation

  1. 1. Predicng  Reassignments  of  Bug  Reports  —  an  Exploratory  InvesgaonAhmed Lamkanfi, Serge DemeyerAnsymoAntwerp Systems and Software ModellingProceedings of the 17th European Conference on Software Maintenance andReengineeringWednesday 12 June 2013
  2. 2. DevelopersEnd-usersWednesday 12 June 2013
  3. 3. Report through the bug-tracking systemDevelopersEnd-usersWednesday 12 June 2013
  4. 4. Limited technical backgroundReport through the bug-tracking systemDevelopersEnd-usersWednesday 12 June 2013
  5. 5. Provides vital informationLimited technical backgroundReport through the bug-tracking systemDevelopersEnd-usersWednesday 12 June 2013
  6. 6. Provides vital informationLimited technical backgroundReport through the bug-tracking systemDevelopersEnd-usersWednesday 12 June 2013
  7. 7. Provides vital informationLimited technical backgroundReport through the bug-tracking systemDevelopersEnd-usersPotential incorrect informationWednesday 12 June 2013
  8. 8. Provides vital informationLimited technical backgroundReport through the bug-tracking systemDevelopersEnd-usersComponent attributePotential incorrect informationWednesday 12 June 2013
  9. 9. /10Product JDTComponentCoreSeverity NormalSummary“Dangerous Eclipsebehavior on file savewhen out of disk space”3Wednesday 12 June 2013
  10. 10. /10Product JDTComponentCoreSeverity NormalSummary“Dangerous Eclipsebehavior on file savewhen out of disk space”Product PlatformComponent TextSeverity CriticalSummary“[implementation]DangerousEclipse behavior on filesave when out of diskspace”3Wednesday 12 June 2013
  11. 11. /10Product No. of Bugs No. of ComponentsPlatform 35.378 21PDE 7.136 5JDT 16.472 6CDT 7.750 19GEF 852 7Core 54.999 116Bugzilla 5.274 21Firefox 67.348 28Thunderbird 18.513 18SeaMonkey 7.808 33Cases  under  Study:  Eclipse  and  Mozilla4Wednesday 12 June 2013
  12. 12. /10Product No. of Bugs No. of ComponentsPlatform 35.378 21PDE 7.136 5JDT 16.472 6CDT 7.750 19GEF 852 7Core 54.999 116Bugzilla 5.274 21Firefox 67.348 28Thunderbird 18.513 18SeaMonkey 7.808 33Cases  under  Study:  Eclipse  and  Mozilla4Some products withlarge number of componentsWednesday 12 June 2013
  13. 13. /10Cases  under  Study:  Eclipse  and  MozillaProduct No. of Reassigned % of ReassignedPlatform 8326 23.5PDE 1290 18.1JDT 4952 30.1CDT 2328 30.0GEF 71 8.3Core 17999 32.7Bugzilla 489 9.3Firefox 10474 15.6Thunderbird 1697 9.2SeaMonkey 1224 15.75Wednesday 12 June 2013
  14. 14. /10Cases  under  Study:  Eclipse  and  MozillaProduct No. of Reassigned % of ReassignedPlatform 8326 23.5PDE 1290 18.1JDT 4952 30.1CDT 2328 30.0GEF 71 8.3Core 17999 32.7Bugzilla 489 9.3Firefox 10474 15.6Thunderbird 1697 9.2SeaMonkey 1224 15.75✓Many reassignments✓Difference between Eclipse and MozillaWednesday 12 June 2013
  15. 15. /101. Many reports are assigned to wrong component6Wednesday 12 June 2013
  16. 16. /101. Many reports are assigned to wrong component2. More bug triaging steps needed6Wednesday 12 June 2013
  17. 17. /101. Many reports are assigned to wrong component2. More bug triaging steps needed3. Decrease of productivity6Wednesday 12 June 2013
  18. 18. /101. Many reports are assigned to wrong component2. More bug triaging steps needed3. Decrease of productivity✓ Tool support to aid bug triagers is useful6Wednesday 12 June 2013
  19. 19. /10History of all reported bugsUncover/mine facts about historyMake predictions about futureBug Database7Wednesday 12 June 2013
  20. 20. /10Can we predict which bugs will be assigned to adifferent component?8Wednesday 12 June 2013
  21. 21. /10Can we predict which bugs will be assigned to adifferent component?Binary classification problem8Wednesday 12 June 2013
  22. 22. /10Can we predict which bugs will be assigned to adifferent component?Binary classification problemSummary and initial component as input8Wednesday 12 June 2013
  23. 23. /10Can we predict which bugs will be assigned to adifferent component?Binary classification problemSummary and initial component as inputEach product with separate classifier8Wednesday 12 June 2013
  24. 24. /10Can we predict which bugs will be assigned to adifferent component?Binary classification problemSummary and initial component as inputPrecision and Recall of predictionsEach product with separate classifier8Wednesday 12 June 2013
  25. 25. /10Can we predict which bugs will be assigned to adifferent component?Binary classification problemEvaluation using 10-fold cross-validationSummary and initial component as inputPrecision and Recall of predictionsEach product with separate classifier8Wednesday 12 June 2013
  26. 26. /10Accuracy  ResultsReassignedReassigned No ReassignmentNo ReassignmentProduct Precision Recall Precision RecallPlatform 0.58 0.54 0.88 0.90PDE 0.62 0.55 0.92 0.94JDT 0.60 0.54 0.82 0.85CDT 0.93 0.82 0.93 0.98GEF 0.86 0.30 0.95 1.00Core 0.80 0.84 0.94 0.92Bugzilla 0.74 0.26 0.94 0.99Firefox 0.70 0.58 0.94 0.96Thunderbird 0.44 0.08 0.93 0.99SeaMonkey 0.92 0.42 0.92 1.009Wednesday 12 June 2013
  27. 27. /10Accuracy  ResultsReassignedReassigned No ReassignmentNo ReassignmentProduct Precision Recall Precision RecallPlatform 0.58 0.54 0.88 0.90PDE 0.62 0.55 0.92 0.94JDT 0.60 0.54 0.82 0.85CDT 0.93 0.82 0.93 0.98GEF 0.86 0.30 0.95 1.00Core 0.80 0.84 0.94 0.92Bugzilla 0.74 0.26 0.94 0.99Firefox 0.70 0.58 0.94 0.96Thunderbird 0.44 0.08 0.93 0.99SeaMonkey 0.92 0.42 0.92 1.009Wednesday 12 June 2013
  28. 28. /10Conclusions10Wednesday 12 June 2013
  29. 29. /10Conclusions10Bugs are often assigned to the wrongcomponentWednesday 12 June 2013
  30. 30. /10Conclusions10Bugs are often assigned to the wrongcomponentOpportunity to improve bug triaging processby predicting component reassignmentsWednesday 12 June 2013

×