IITP: A Supervised Approach for Disorder Mention Detection and Disambiguation

Proceedings of the 8th International Workshop on Semantic Evaluation (SemEval 2014), pages 314–318, Dublin, Ireland, August 23-24, 2014. IITP: A Supervised Approach for Disorder Mention Detection and Disambiguation Utpal Kumar Sikdar, Asif Ekbal and Sriparna Saha Department of Computer Science and Engineering Indian Institute of Technology Patna, India {utpal.sikdar,asif,sriparna}@iitp.ac.in Abstract In this paper we briefly describe our super- vised machine learning approach for dis- order mention detection system that we submitted as part of our participation in the SemEval-2014 Shared task. The main goal of this task is to build a system that automatically identifies mentions of clini- cal conditions from the clinical texts. The main challenge lies due in the fact that the same mention of concept may be repre- sented in many surface forms. We develop the system based on the supervised ma- chine learning algorithms, namely Condi- tional Random Field and Support Vector Machine. One appealing characteristics of our system is that most of the features for learning are extracted automatically from the given training or test datasets with- out using deep domain specific resources and/or tools. We submitted three runs, and best performing system is based on Condi- tional Random Field. For task A, it shows the precision, recall and F-measure values of 50.00%, 47.90% and 48.90%, respec- tively under the strict matching criterion. When the matching criterion is relaxed, it shows the precision, recall and F-measure of 81.50%, 79.70% and 80.60%, respec- tively. For task B, we obtain the accuracies of 33.30% and 69.60% for the relaxed and strict matches, respectively. 1 Introduction The SemEval-2014 Shared Task 7 is concerned with the analysis of clinical texts, particularly for disorder mention detection and disambiguation. This work is licensed under a Creative Commons At- tribution 4.0 International Licence. Page numbers and pro- ceedings footer are added by the organisers. Licence details: http://creativecommons.org/licenses/by/4.0/ The purpose of this task is to enhance current research in Natural Language Processing (NLP) methods used in the clinical domain. The task is a continuation of the CLEF/eHealth ShARe 2013 Shared Task. In particular there were two specific tasks, viz. (i). Task A: To identify disorder men- tions from biomedicine domain and (ii) Task B: To classify each mention with respect to the Uni- fied Medical Language System (UMLS) Concept Unique Identifier (CUI). The task is challenging in the sense that the same mention of concept may be represented in many surface forms and men- tion may appear in the different parts of texts. Some systems (Cogley et al., 2013; Zuccon et al., 2013; Tang et al., 2013; Cogley et al., 2013) are available for disorder mention detection. Look- ing at the challenges and resources available at our hand we planned to adapt our existing system (Sikdar et al., 2013) for disorder mention detec- tion. The original architecture was conceptualized as part of our participation in the BioCreative-IV Track-2 Shared Task on Chemical Compound and Drug Name Recognition. Although our submit- ted system for SemEval-14 shared task is in line with BioCreative-IV1, it has …

Content truncated to 500 words