Table 5.

Comparison of models built on different libraries for SAS threshold of 3.5 Å

λ1	400 (12)	600 (10)	600 (9)	250 (7)	200 (6)	100 (5)
0	0.76	0.76	0.76	0.74	0.69	0.72
0.1	0.77	0.77	0.77	0.75	0.7	0.72
0.2	0.78	0.77	0.77	0.75	0.71	0.72
0.3	0.78	0.78	0.78	0.76	0.72	0.73
0.4	0.79	0.78	0.79	0.76	0.73	0.73
0.5	0.79	0.79	0.79	0.76	0.74	0.73
0.6	0.79	0.8	0.8	0.77	0.75	0.74
0.7	0.8	0.8	0.8	0.77	0.75	0.74
0.8	0.8	0.8	0.8	0.78	0.76	0.74
0.9	0.79	0.79	0.8	0.77	0.76	0.75
1	0.77	0.77	0.78	0.76	0.76	0.74

Here each library is denoted as X(Y), where X is the number of fragments in the library, each of length Y. The ranking performance of a given multi-viewpoint IR model for a given library is given in terms of AUC. The multi-viewpoint model contains LDA model with weight λ₁ and TF vector space model with weight 1 − λ₁.