In this article we have discussed the use of data fusion methods to combine the rankings resulting from similarity searches of chemical datasets. Our experiments, which have employed a range of types of molecule and performance criterion, demonstrate that use of a fusion rule such as SUM will generally result in a level of performance (however this is quantified) that is at least as good (when averaged over a number of searches) as the best individual measure: since the latter often varies from one target structure to another in an unpredictable manner, the use of a fusion rule will generally provide a more consistent level of searching performance than if just a single similarity measure is available.

0 0

Post a comment