• Fulltext


        Click here to view fulltext PDF

      Permanent link:

    • Keywords


      Biomedical event extraction; biomedical triggers; machine learning; CapsNet.

    • Abstract


      Biomedical Event Extraction (BEE) is a demanding and prominent technology that attracts the researchers and scientists in the field of natural language processing (NLP). The conventional method relies mostly on external NLP packages and manual designed features, where the features engineering is complex and large. In addition, the conventional methods on BEE uses a pipeline process that splits a task into many subtasks, however, the relationship between these sub-tasks is not defined. In this paper, such limitations are avoided using the combination technique that relies on Capsule Network (CapsNet) to perform a task. The CapsNet is used for the extraction of feature representation from the input corpora and then the combination technique reconstructs the events from RNN output. This method extracts the tasks from a BEE over several annotated corpora that extract the events from the molecular level in case of multi-level events. The proposed model is compared with state-of-the-art models over various text corpora datasets. The results show an improved rate of accuracy of CapsNet classification over cancer biomedical events than the existing methods.

    • Author Affiliations



      1. Department of CSE, Sri Ramakrishna Institute of Technology, Coimbatore, India
      2. Department of CSE, SNS College of Technology, Coimbatore, India
      3. Department of ECE, Karpagam College of Engineering, Coimbatore, India
      4. Business Information Technology Division, Department of Statistics, Faculty of Commerce and Accountancy, Chulalongkorn University, Bangkok, Thailand
    • Dates

  • Sadhana | News

    • Editorial Note on Continuous Article Publication

      Posted on July 25, 2019

      Click here for Editorial Note on CAP Mode

© 2022-2023 Indian Academy of Sciences, Bengaluru.