 Poster presentationExploring benchmark dataset bias in ligand based virtual screeningKnut Baumann and SG Rohrer Institute of Pharmaceutical Chemistry, Braunschweig University of Technology, Beethovenstr. 55, 38106 Braunschweig, Germany  corresponding author email
from 3rd German Conference on Chemoinformatics Goslar, Germany. 11-13 November 2007 Chemistry Central Journal 2008,
2(Suppl 1):P1doi:10.1186/1752-153X-2-S1-P1 First paragraph (this article has no abstract)
A common finding of many reports evaluating VS methods is that validation results vary considerably with changing datasets, i.e. chemical space of the active ligands. It is assumed that these dataset specific effects are caused by the self-similarity and cluster structure inherent to these datasets. | 
on Google Scholar
