郭展宏

Chan-Hung Kuo

碩士論文 (2010)

以概念為基之資訊檢索機制研發

Development of a Concept-based Information Retrieval Mechanism

關鍵字 Keywords

資訊檢索系統, 概念地圖, 資訊需求, 領域不熟習者

Information retrieval, Concept map, Information needs, Layman

摘要

隨著資訊內容急遽的增加，在大量的資訊內容中找尋資料並滿足資訊需求是項艱難的任務。現今之資訊檢索系統可協助使用者獲取符合其資訊需求的內容，但當使用者對領域不熟悉以致無法提供合適查詢主題(Search Items)時，資訊檢索的時效與正確率將大幅降低。針對上述問題，本研究提出以概念為基之資訊檢索機制(Concept-based Information Retrieval Mechanism，CBIR)，透過與使用者之互動，解析使用者資訊需求並轉換為概念地圖(Concept Map)，再藉由概念地圖搜尋與擷取符合使用者需求之資訊。實驗結果顯示，針對領域不熟習者(Layman)本機制優於傳統之Vector Space Model (VSM)檢索模式，可有效輔助使用者進行概念組織，找出合適的查詢主題，以提升檢索的正確率與時效。

Abstract

As the amount of electronic data is increasing, it is getting harder to search information effectively. Contemporary information retrieval systems do help users to obtain required information, however, the effectiveness and precision of information retrieval may remarkably decrease especially when the users are layman of a targeted domain. This thesis proposed a Concept-based Information Retrieval Mechanism (CBIR) to help users retrieve information fulfilling their information needs through interactive information requirement analysis. The information requirement analysis process helps users organize concepts, which are gradually transferred into a concept map used for information retrieval. Experimental results showed that proposed Concept-based Information Retrieval Mechanism is more effective than traditional information retrieval mechanism (Vector Space Model). The research result may be used for user concept organization to efficiently obtain required information as well as enhance information retrieval precision.