Title

Creation Assistant for Easy Assignment

Presenter Information

Hank Burton
Marshall Brooks
Justin Canada

Document Type

Oral Presentation

Location

SURC 140

Start Date

16-5-2013

End Date

16-5-2013

Abstract

Software development projects receive many bug reports each day. Each of these reports needs to be examined and decisions made about how to handle the report. This process is called bug report triage. One decision that is frequently made is to which software developer to assign the bug report. There have been many efforts toward automating this decision, with the most promising approaches using machine learning algorithms. However, creating a bug report assignment recommender using machine learning is a complex process that must be tailored to each software development project. This project presents a tool, called the Creation Assistant for Easy Assignment (CASEA), which assists in creating a bug report assignment recommender for a software development project. CASEA uses data mining to pull reports from a Bugzilla bug repository via XML-RPC, assists in creating heuristics to know who fixed a bug, helps filter the data to recommend only current project developers, and creates a bug report assignment recommender using the SVM machine learning algorithm. Feedback on the effectiveness of the created recommender is provided using precision and recall metrics. The user can then adjust the filtering and heuristics until they are satisfied with the recommender performance. We evaluated CASEA by creating a recommender for the Eclipse IDE project and found that we could create an assignment recommender within 10 percent of the precision and recall of a hand tailored recommender. This software makes using a bug report assignment recommenders practical, potentially saving software development companies both time and money.

Faculty Mentor(s)

John Anvik

Additional Mentoring Department

Computer Science

This document is currently not available here.

Share

COinS
 
May 16th, 8:20 AM May 16th, 8:40 AM

Creation Assistant for Easy Assignment

SURC 140

Software development projects receive many bug reports each day. Each of these reports needs to be examined and decisions made about how to handle the report. This process is called bug report triage. One decision that is frequently made is to which software developer to assign the bug report. There have been many efforts toward automating this decision, with the most promising approaches using machine learning algorithms. However, creating a bug report assignment recommender using machine learning is a complex process that must be tailored to each software development project. This project presents a tool, called the Creation Assistant for Easy Assignment (CASEA), which assists in creating a bug report assignment recommender for a software development project. CASEA uses data mining to pull reports from a Bugzilla bug repository via XML-RPC, assists in creating heuristics to know who fixed a bug, helps filter the data to recommend only current project developers, and creates a bug report assignment recommender using the SVM machine learning algorithm. Feedback on the effectiveness of the created recommender is provided using precision and recall metrics. The user can then adjust the filtering and heuristics until they are satisfied with the recommender performance. We evaluated CASEA by creating a recommender for the Eclipse IDE project and found that we could create an assignment recommender within 10 percent of the precision and recall of a hand tailored recommender. This software makes using a bug report assignment recommenders practical, potentially saving software development companies both time and money.