Knowledge emergence from biomedical text using the link grammar and metamap

Abstract

There have been numerous examples of application of NLP techniques to extract PPI data from natural language texts, but few for other purposes. Most of the previously developed methods performed only extraction of information with no further analysis or inference. With the advancement in biomedical science, it has become imperative to extract and then combine information from multiple disjoint researches, studies and articles to infer new hypotheses, and expand knowledge. We developed a method for extracting relationships using Link Grammar Parser while employing MetaMap as a named entity recognizer. The rules created from our “Tagger” were fed to the extractor which performed the main extraction task. When applied to MEDLINE abstracts, the system was able to extract relevant relationships with good precision and recall. Afterwards, the extracted data is used for knowledge emergence by combining multiple pieces of information to infer new knowledge using our proposed similarity measure. Such system can be used to provide new insights into the actions of drugs and other substances.