Computer-aided Knowledge Search and Extraction

[[abstract]]以電腦來輔助知識管理是組織提昇競爭優勢的關鍵手段之一。Holsapple將知識依其功能性分為程序性知識、推論性知識、及描述性知識三種類型，其中又以描述性知識最為變化多端，因此在搜尋與擷取時特別困難。Ju and Sun從電腦輔助知識管理的角度提出了一個知識模式，對資料、資訊、和知識之間的轉變及共通性的編碼原則作了詳細的闡述，同時對知識的詮釋也有概略的說明。Ju and Sun的論述激發了本研究對搜尋與擷取描述性知識的創新想法。本研究以該知識模式為基礎並參考相關文獻，從而勾勒出知識搜尋與擷取的環境以便設計合用的知識物件搜尋與擷取機制，同時以主題地圖為核心技術，設計出管制詞彙編輯器及後設資訊建構器，讓知識工程師及知識創作者能方便地編輯管制詞彙及有助於知識搜尋與擷取的後設資訊，也設計出知識物件搜尋引擎及知識物件導覽器，以達成既不泛濫又無遺珠之憾的高精準度知識搜尋與擷取。最後本研究小規模地評估了研究成果，結論是整個構想一樣可以運用在程序性知識和推論性知識上，值得知識管理系統設計者參考。嚴謹的效率和效能評估則有待進一步的努力。[[abstract]]Computer-aided knowledge management is one of the critical means to enhance organizational competitiveness. Holsapple classifies knowledge according to its function into procedural knowledge, inferential knowledge, and descriptive knowledge. Among these three kinds of knowledge, descriptive knowledge is the most versatile, and therefore is the most difficult to search and extract. Ju and Sun propose a knowledge model for knowledge management, in which the transition between data, information, and knowledge is detailed, and a common codification principle is described. Ju and Sun’s proposition inspired us to pursue an innovative approach to the search and extraction of descriptive knowledge. This study is based on the knowledge model and review of relevant literature. It has characterizes the environment of knowledge search and extraction for designing a suitable knowledge object search mechanism. It used Topic Maps as core technology and designed a controlled-vocabulary editor and a meta-infrmation constructor to help knowledge engineers and knowledge authors edit the controlled-vocabulary and construct meta-information that support knowledge search and extraction. It also designed a knowledge object search engine and a knowledge object navigator to perform high-precision search and extraction, so that nothing is too much and nothing is missed. With a small-scaled evaluation we find that the result are equally applicable to procedural knowledge and inferential knowledge, and therefore is worth the attention of knowledge management systems developers. A formal assessment of efficiency and effectiveness of the proposed scheme, however, is for further studies