Table 3.

Comparing the average AUC for various Multi-viewpoint IR methods

λ1	SAS=2			SAS=3.5			SAS=5
	I	II	III	I	II	III	I	II	III
0	0.89	0.87	0.8	0.77	0.72	0.64	0.75	0.73	0.68
0.1	0.89	0.87	0.81	0.78	0.73	0.65	0.75	0.73	0.69
0.2	0.9	0.88	0.81	0.78	0.74	0.66	0.75	0.73	0.69
0.3	0.9	0.88	0.82	0.79	0.75	0.67	0.75	0.74	0.7
0.4	0.91	0.89	0.83	0.79	0.76	0.69	0.77	0.74	0.7
0.5	0.91	0.9	0.85	0.8	0.77	0.7	0.75	0.74	0.71
0.6	0.91	0.9	0.86	0.8	0.78	0.72	0.75	0.73	0.71
0.7	0.91	0.9	0.88	0.8	0.78	0.75	0.75	0.73	0.71
0.8	0.91	0.91	0.89	0.8	0.79	0.77	0.74	0.72	0.71
0.9	0.91	0.9	0.9	0.8	0.78	0.77	0.72	0.7	0.7
1	0.9	0.9	0.9	0.78	0.78	0.78	0.68	0.68	0.68

The multi-viewpoint models are obtained by combining latent dirichilet allocation topic model (LDA) with weight λ₁ and one of the following vector space models (i) term frequency (TF) (I), (ii) term frequency inverse document frequency (TF-IDF) (II) and (iii) boolean (BOOL) (III) with weight λ₂. Since λ₂=1− λ₁, we have not mentioned their values explicitly in the table.