Concept-based medical document retrieval: THCIB at CLEF eHealth lab 2013 task 3

Xiaoshi Zhong, Yunqing Xia, Zhongda Xie, Sen Na, Qinan Hu, Yaohai Huang

Research output: Contribution to journalConference articlepeer-review

4 Citations (Scopus)

Abstract

We describe our participation in the task 3 of ShARe/CLEF eHealth Lab 2013: information retrieval to address questions patient may have when reading clinical reports. In our experiments, we focus mainly on two levels of analysis, namely query analysis and document analysis, to disclose the relevance between query and documents. In terms of query analysis, we first observe each medical-oriented query to find its identical or related UMLS concepts derived from the query, which may help to induce relevant results that refer to the same thing but are represented in other surface forms. In such manner, we extend the query based on the medical concepts so as to achieve a bigger coverage. In terms of document analysis, we leverage different scores (e.g., relevance score, PageRank score, HITS score and layout score) as feature to re-rank the documents of search results. With those two levels of analysis, we implement a concept-based method and a topic-based method to accomplish the task of medical document retrieval. Experiments indicate that the proposed method is effective.

Original languageEnglish
JournalCEUR Workshop Proceedings
Volume1179
Publication statusPublished - 2013
Externally publishedYes
Event2013 Cross Language Evaluation Forum Conference, CLEF 2013 - Valencia, Spain
Duration: 23 Sept 201326 Sept 2013

Keywords

  • Concept
  • Document ranking
  • Medical document retrieval
  • Query analysis
  • Topic

Fingerprint

Dive into the research topics of 'Concept-based medical document retrieval: THCIB at CLEF eHealth lab 2013 task 3'. Together they form a unique fingerprint.

Cite this