結合資訊檢索與分群演算法建構知識地圖

Combining Information Retrieval and Clustering Algorithm to Construct an Knowledge Map

邱登裕(Deng-Yiv Chiu) ； 潘雅真(Ya-Chen Pan)

《資訊管理學報》 13卷S期 (2006/10) Pp. 137-160

https://doi.org/10.6382/JIM.200610.0007

摘要

知識地圖(Knowledge Map, K-Map)是用來呈現知識分布的其中一個方法。在建立知識地圖之前，必須分析文件的內容，擷取其重要內容並建立關聯程度。本研究首先利用中研院所提供之CKIP中文斷詞系統，將文件內容進行斷詞及詞性標註，透過資訊檢索(Information Retrieval)、資料探勘(Data Mining)及分析等技術，將所擷取之重要特徵詞給予其代表性之權重，並進行相似度的計算與分群，最後將分析後的特徵詞與文件對應到知識地圖中的概念。
本研究實作了一個知識地圖系統，使用者可以籍由知識地圖的呈現方式，快速的找尋所需知識，並可更進一步觀察其相關知識的分布，促進組織內部知識的保留與共享，達到知識管理的目標。

關鍵字

知識地圖 ； 資訊檢索 ； 資料探勘

並列摘要

Knowledge map is one kind of technique to represent knowledge. Before building up a knowledge map, it is necessary to analyze related document. The two main purpose of this research are to extract the important content and to create the association level. First, CKIP system is used to identify Chinese words and to tag the morphological features. Secondly, techniques of information retrieval and data mining and analysis are used to weight importance of extracted keywords. Thirdly, similarities between documents are computed and similar documents are grouped together. Finally, extracted keyword and document are mapped to the concept of the knowledge map.
The research presents a Knowledge Map system. Through KMS, users can perform fast knowledge search and moreover they can observe the distribution of related knowledge. KMS can also be used to aid an enterprise in retaining and sharing inter-organizational knowledge to achieve the goal of knowledge management.

並列關鍵字

Knowledge Map ； Information Retrieval ； Data Mining

參考文獻

CKIP中文斷詞系統

(1993)。中文詞類分析。

孫振凱(2001)。利用網頁建構知識分布圖(碩士論文)。國立中山大學資訊管理研究所。

Baker, L. D.,McCallum, A. K.(1998).Distributional clustering of words for text categorization.Proceeding of ACM SIGIR international conference on information retrieval.(Proceeding of ACM SIGIR international conference on information retrieval).

Davenport, T.,De Long, D.,Beers, M.(1998).Successful knowledge projects.sloan management review.39,43-57.