整合查詢擴展融合與MeSH醫學字詞重排序之醫學文件檢索方法;A medical document retrieval method of integrating the query expansion fusion technique with the MeSH-based re-ranking

NCUIR > School of Management at National Central University > Graduate Institute of Information Management > Electronic Thesis & Dissertation > Item 987654321/77646

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/77646

Title:	整合查詢擴展融合與MeSH醫學字詞重排序之醫學文件檢索方法;A medical document retrieval method of integrating the query expansion fusion technique with the MeSH-based re-ranking
Authors:	胡馨如;Hu, Sin-Ru
Contributors:	資訊管理學系
Keywords:	資訊檢索;醫學文件檢索;MeSH;查詢擴展;文件重排序;Information Retrieval;Medical Document Retrieval;MeSH (Medical Subject Headings);Query Expansion;Document Re-ranking
Date:	2018-07-30
Issue Date:	2018-08-31 14:51:38 (UTC+8)
Publisher:	國立中央大學
Abstract:	資訊及醫療技術的快速進步，存在於網際網路上的數位醫學文件也與日俱增，如何管理以及搜尋這些文件成為重要的議題。而使用者搜尋時給予的資訊有限，無法有效涵蓋搜索條件，為此過去學者已提出許多查詢擴展技術來解決此問題。而本研究欲發展一個融合多種查詢擴展之效能融合方法(Efficacy Fusion, E.F.)，E.F.方法使用查詢擴展的檢索效能分配融合權重，並且比起單一查詢擴展方法，融合多種查詢擴展方法更能幫助使用者找出相關的文件。此外，有別於過去研究學者所提出將MeSH樹結構視為一種查詢擴展應用，本研究應用MeSH樹結構將檢索出的相關文件進行重排序，此方法能避免使用MeSH樹結構作為查詢擴展應用時所發生的噪音問題(找到許多不相關的文件)，並加快讓使用者找到最相關的文件。本研究所使用之資料集為TREC 2007 Genomics資料集，並利用資料集提供之評估指標進行方法的檢索效能驗證。最終實驗結果顯示，本研究所提出之整合查詢擴展融合與MeSH醫學字詞重排序方法能提升檢索效能，且檢索效能改進幅度高於過去研究者所提出之方法。;Due to the development of information and medical technology, digital medical documents on the Internet have been growing rapidly. How to manage and search these documents has become an important issue. However, the information given by users for search is limited and usually cannot effectively cover the conditions of information requirement. For this reason, scholars have proposed many query expansion techniques to solve this problem. This study aims to develop a method “Efficacy Fusion” that integrates multiple query expansions. Efficacy Fusion method use the retrieval performance of query expansion as the weight of combination. This method can be appropriately assigned weights, moreover, the combination of the different query expansion methods is more efficient than the use of any of them separately. In addition, this study used the MeSH tree structure to re-rank the retrieved related document. It can avoid a noise issue while expanding terms by MeSH tree structure and let users find the most relevant documents faster. This study used TREC 2007 Genomics dataset for experiment. MAP for performance evaluation. The experimental results show that the method proposed in this study can improve the retrieval performance and outperform past study.
Appears in Collections:	[Graduate Institute of Information Management] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	278	View/Open

社群 sharing

Loading...