A Semi-Supervised Framework for MMMs-Induced Fuzzy Co-Clustering with Virtual Samples

Table 4

Comparison of recognition rates with and without virtual objects (CiteSeer-training).

Number of virtual objects per supervised object
	0.01	0.05	0.1	0.2
No virtual objects	0.542
1 (GBD)	0.551	0.551	0.548	0.546
1 (GBA)	0.552	0.545	0.549	0.543
2 (GBD)	0.554	0.550	0.554	0.550
2 (GBA)	0.553	0.550	0.548	0.546
2 (GBD+GBA)	0.554	0.554	0.547	0.549
4 (GBD+GBA)	0.541	0.542	0.534	0.533
5 (GBD)	0.534	0.536	0.532	0.544
5 (GBA)	0.504	0.530	0.527	0.532