Cross-Project Defect Prediction Using Transfer Learning with Long Short-Term Memory Networks

Table 5

f-measure of a set of CPDP for different models when Xerces as the target project.

Source project	LR	NNFilter	TCA	DBN	DPDBN	DPCNN	DPTCNN	TLSTM	DPTLSTM
Camel-1.6	0.517	0.4	0.562	0.559	0.514	0.575	0.594	0.747	0.729
Forrest-0.8	0.336	0.355	0.332	0.418	0.347	0.487	0.482	0.549	0.415
Log4j-1.2	0.708	0.734	0.712	0.634	0.729	0.643	0.634	0.676	0.762
Synapse-1.2	0.579	0.508	0.692	0.519	0.542	0.702	0.73	0.749	0.774
Xalan-2.7	0.782	0.78	0.779	0.739	0.763	0.764	0.773	0.643	0.71

The values in bold represent the most optimal values in a row of data.