tailieunhanh - Coreference resolution in Vietnamese electronic medical records

This paper tackles the problem of coreference resolution in Vietnamese EMRs. Unlike in English ones, in Vietnamese clinical texts, verbs are often used to describe disease symptoms. So we first define rules to annotate verbs as mentions and consider coreference between verbs and other noun or adjective mentions possible. | VNU Journal of Science: Comp. Science & Com. Eng., Vol. 34, No. 2 (2018) 33–43 Coreference Resolution in Vietnamese Electronic Medical Records Hung D. Nguyen1,∗, Tru H. Cao2 1 Faculty 2 Faculty of Information Technology, Monash University, Victoria, Australia of Computer Science and Engineering, Ho Chi Minh University of Technology, Ho Chi Minh City, Vietnam Abstract Electronic medical records (EMR) have emerged as an important source of data for research in medicine and information technology, as they contain much of valuable human medical knowledge in healthcare and patient treatment. This paper tackles the problem of coreference resolution in Vietnamese EMRs. Unlike in English ones, in Vietnamese clinical texts, verbs are often used to describe disease symptoms. So we first define rules to annotate verbs as mentions and consider coreference between verbs and other noun or adjective mentions possible. Then we propose a support vector machine classifier on bag-of-words vector representation of mentions that takes into account the special characteristics of Vietnamese language to resolve their coreference. The achieved F1 score on our dataset of real Vietnamese EMRs provided by a hospital in Ho Chi Minh city is . To the best of our knowledge, this is the first research work in coreference resolution on Vietnamese clinical texts. Received 15 August 2018, Revised 16 November 2018, Accepted 25 December 2018 Keywords: Clinical text, support vector machine, bag-of-words vector, lexical similarity, unrestricted coreference. 1. Introduction NLP community for the last 20 years. In the early days, the focus was primarily put on the general domain of mostly newswire corpora. Firstly approached with hand-crafted methods using discourse theories such as focusing or centering [1, 2], coreference resolution received the first learning-based treatment by Connolly et al. in 1994 [3] that casted it as a classification problem. Since then, several supervised models have .