An Improved Method for Cross-Project Defect Prediction by Simplifying Training Data

Table 2

Data statistics of the projects used in our experiments.

Repository	Project	Version	#Instance	#Defect	% Defect
PROMISE	Ant	1.7	745	166	22.3%
	Camel	1.6	965	188	19.5%
	Ivy	2.0	352	40	11.4%
	Jedit	3.2	272	90	33.1%
	Lucene	2.4	340	203	59.7%
	Poi	3.0	442	281	63.6%
	Synapse	1.2	256	86	33.6%
	Velocity	1.4	196	147	75.0%
	Xalan	2.6	885	411	46.4%
	Xerces	1.4	588	437	74.3%
AEEEM	Equinox	1.1.2005–6.25.2008	324	129	39.8%
	Eclipse JDT core (Eclipse)	1.1.2005–6.17.2008	997	206	20.7%
	Apache Lucene (Lucene2)	1.1.2005–10.8.2008	692	20	2.9%
	Mylyn	1.17.2005–3.17.2009	1,862	245	13.2%
	Eclipse PDE UI (Pde)	1.1.2005–9.11.2008	1,497	209	14.0%