Constructing biological knowledge base using named entities recognition and term collocation

• Main Authors: Supattanawaree Thipcharoen, Watshara Shoombuatong, Samerkae Somhom, Rattasit Sukhahuta, Jeerayut Chaijaruwanich
• 格式: 雜誌
• 出版: 2018
• 主題: Biochemistry, Genetics and Molecular Biology; Chemistry; Materials Science; Mathematics; Physics and Astronomy
• 在線閱讀: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84961817052&origin=inward; http://cmuir.cmu.ac.th/jspui/handle/6653943832/55292

© 2016, Chiang Mai Journal of Science. All rights reserved. Over the last few decades, the publishing of biological literature has dramatically increased due to technological developments. Thus, a crucial process is to extract information from this large number of writings by utilizing a biological named entity (NER) approach to automatically label corresponding biological terms. It is desirable to propose an effective method to identify biological named entities and automatically establish the specific knowledge base from biological literature. Herein, we made efforts in investigating biological information extraction for establishing specific knowledge as follows: 1) proposing NER method based on the efficient conditional random fields (CRFs) model, called NER-CRF, for performing on the benchmarking data (JNLPBA2004). The proposed NER method provided a higher result with 90.42% recall, 97.74% precision, and 94.30% F-measure, compared with the existing method with 75.99% recall, 69.42% precision, and 72.55% F-measure; 2) applying the Poisson approach for constructing an interpretability biological knowledge network to give good understanding to the global properties collocation of biological terms from the literature. Our finding provided the collocations of biological terms from the literature providing some insights to the specific biological literature.