Turbo similarity searching

Fuse the sorted lists SD(0)-SD(k) to give the final output from the turbo similarity search. 

Figure 2. Use of nearest neighbors for turbo similarity searching.

Table 6 Mean recall at 5°/o for Conventional Similarity Searching (SS) with just a Single Reference Structure and Turbo Similarity Searching (TSS) using different numbers of NNs...

Table 8 Mean Recalls at 5% for Similarity Searching (SS), Conventional Turbo Similarity Searching using Croup Fusion (TSS-CF) and Turbo Similarity Searching using Substructural Analysis (TSS-SSA) and Binary Kernel Discrimination (TSS-BKD)...

Gardiner EJ, Gillet VJ, Haranczyk M, Hert J, Holliday JD, Malim N, Patel Y, Willett P (2009) Turbo similarity searching effect of FP and dataset on virtual-screening performance. Stat Anal Data Mining 2 103-114... 

Interestingly, turbo similarity is reminiscent of library search procedirres, where a given query yields a set of hits, each of which is used in a subsequent qnery to broaden the search [107]. 

---