Defect Prediction on Software Projects with Limited Historical Data

PhD Thesis Proposal Defence


Title: "Defect Prediction on Software Projects with Limited Historical Data"

by

Mr. Jaechang NAM


Abstract:

Software defect prediction is one of active research areas in software 
engineering. Researchers have proposed many defect prediction algorithms and 
metrics. However, software defect prediction has a limitation that it is 
difficult to build prediction models on software projects with limited 
historical data such as defect information.

To overcome this limitation, we propose three techniques that can build 
prediction models on projects with limited historical data. First, we adopt a 
state-of-the-art transfer learning technique, transfer component analysis 
(TCA), and propose TCA+ to build a prediction model using other projects. 
Second, we propose cross-domain defect prediction that enables cross-project 
defect prediction on projects with different metric sets. Lastly, we propose 
CLAMI for defect prediction on unlabeled datasets to build a prediction model 
using a project that does not have any defect information.


Date:			Thursday, 29 January 2015

Time:			2:00pm - 4:00pm

Venue:			Room 3494
			lifts 25/26

Committee Members:	Dr. Sunghun Kim (Supervisor)
			Prof. James Kwok (Chairperson)
 		        Prof. Shing-Chi Cheung
			Dr. Charles Zhang


**** ALL are Welcome ****