Evaluating the Impact of Data Transformation Techniques on the Performance and Interpretability of Software Defect Prediction Models

Table 1

Experimental datasets.

Project	Features	Modules	Defects	Ratio (%)
Camel-1.2	20	608	216	36
Ivy-1.1	20	111	63	57
Jedit-3.2	20	272	90	33
Log4j-1.1	20	109	37	34
Lucene-2.0	20	195	91	47
Lucene-2.2	20	247	144	58
Lucene-2.4	20	340	203	60
Poi-1.5	20	237	141	59
Poi-2.5	20	385	248	64
Poi-3.0	20	442	281	64
Xalan-2.5	20	803	387	48
Xalan-2.6	20	885	411	46
Apache	26	194	98	51
Safe	26	56	22	39
ZXing	26	399	118	30
EQ	61	325	129	40