Semantic event extraction from biological texts using a kernel-based method

Abstract

As research into protein and gene interactions continues to produce vast amount of data, concerning to biological event, there is an increasing need to capture these results in structured formats allowing for computational analysis. Although many efforts have been focused to create databases that store this information in computer readable form, populating these sources largely requires a manual process of interpreting and extracting biological event templates from the biological research literature. Being able to efficiently and systematically automate the extraction of biological events from unstructured text, would improve the content of these databases, and provide methods to collect, maintain, interpret, curate, and discover knowledge needed for research or education. Hence, it is important to have an automated extraction system to extract events from biological texts. In this paper, we present an automated information extraction approach, to identify biological events in text. Our approach is based on, identifying event triggers and extracting event participants by using a kernel learner that operates on dependency and semantic information to calculate similarity between feature vectors. © 2014 Springer International Publishing Switzerland.