A semantic approach for mining biological databases

Abstract

A variety of biological databases are currently available to researchers in the XML format. Homology-related querying on such databases presents several challenges, as most available exhaustive mining techniques do not incorporate the semantic relationships inherent to these data collections. This chapter identifies an index-based approach to mining such data and explores the improvement achieved in the quality of query results by the application of genetic algorithms. Our experiments confirm the widely accepted advantages of index and vector-space based model for biological data and specifically, show that the application of genetic algorithms optimizes the search and achieves higher levels of precision and accuracy in heterogeneous databases and faster query execution across all data collections. © 2009 Springer-Verlag Berlin Heidelberg.