Framework for role discovery using transfer learning

Applied Network Science

Table 5 Results of air-traffic networks

Method	T	BR		EU		US
	S	EU	US	BR	US	BR	EU
RolX (Target tuned)		0.267	0.267	0.256	0.496	0.251	0.459
struc2vec (Target tuned)		0.588	0.710	0.551	0.521	0.582	0.508
Proposed method (Target tuned)		0.634	0.733	0.559	0.546	0.598	0.535
Proposed method (Model selection)		0.588	0.664	0.459	0.451	0.434	0.516
Target-only (Target tuned)		0.741		0.622		0.683

S means the source network, and T means the target network; US, EU, and BR indicate the United States, European Union, and Brazil. The target-only method, which is the result of the model trained by the target network data, showed the difficulty of these tasks