期刊名称：International Journal of Computer Information Systems and Industrial Management Applications
印刷版ISSN：2150-7988
电子版ISSN：2150-7988
出版年度：2013
卷号：5
页码：631-641
出版社：Machine Intelligence Research Labs (MIR Labs)
摘要：The main goal of Biomedical Natural Language Processing (BioNLP) is to capture biomedical phenomena from textual data by extracting relevant entities, information and re- lations between biomedical entities (i.e. proteins and genes). In general, in most of the published papers, only binary re- lations were extracted. In a recent past, the focus is shifted towards extracting more complex relations in the form of bio- molecular events that may include several entities or other rela- tions. In this paper we propose an approach that enables event extraction (detection and classification) of relatively complex bio-molecular events. We approach this problem as a super- vised classification problem and use the well-known algorithms, namely Conditional Random Field (CRF) and Support Vector Machine (SVM) as the underlying classifiers. These algorithms make use of statistical and linguistic features those represent various morphological, syntactic and contextual information of the candidate bio-molecular trigger words. Here the outputs of these classifiers are combined using a newly developed ge- netic ensemble technique. The genetic algorithm based ensem- ble technique will be able to automatically determine the ap- propriate weights of votes for each classifier for each output class in order to combine the outputs of different classifiers us- ing weighted voting. Experiments on the BioNLP 2009 shared task datasets yield the overall average recall, precision and F- measure values of 53.56%, 51.47%, and 52.50%, respectively, on development data.
关键词：Event extraction; Support vector machine; Condi- ; tional random field; Biomedical natural language processing; De- ; tection and classification; Text mining
