- 學位論文
知識分享平台上自動化專家辨識之研究
Automatical Expert Identification on the Knowledge Sharing Platform
摘要
網路社群在現今網路世界中，扮演了很重要的知識分享的角色，社群中的每一位成員可以將自己的想法、經驗或專業知識透過文章發表分享於網路上，知識分享因而更為方便、快速。但由於知識分享平台具有公開性，無論該使用者的知識是否完整正確，皆可以在平台中發表知識文件，導致其他無能力判斷其正確性的使用者可能會吸收到錯誤的知識。在知識分享過程中，一個組織裡必定具有正確專業知識的成員，且該成員在解決問題的能力上也較他人來得優秀，我們可稱之為專家。專家所分享的知識，不但可信度高，且知識的表達較為完整，此類專家能夠幫助組織中其他成員知識的學習之外，也能推動組織正向的發展。因此如何能夠在網路社群中辨識哪一位成員是所謂的專家，便是一個重要的研究議題。本研究期望能夠藉由分析使用者的文章，並透過其他使用者依照自己的知識能力來做的實用性及品質性評價分數，讓系統可以自動評比各使用者的專業程度，並找出足以稱作專家的使用者。本研究所提出的計量導向專家排序，以PageRank及ExpertRank演算法為核心概念，加入計算文章的內容長度，使自動化專家辨識結果更為精確。
並列摘要
The cyber community has played an important role for knowledge sharing in internet. By posting articles in the cyber community, the members can share knowledge with each other directly and conveniently. However, due to the community members are not professional enough, the qualities of the articles in the cyber community are not all good. Therefore, the articles evaluation mechanism is widely applied in cyber communities, and members can refer to the evaluation score before they want to read an article. In the cyber community, people would like to ask someone to help to solve some problems. It is valuable that if the community platform could help to identify who the real expert is in a specific domain. This mechanism would help to enhance the effectiveness of knowledge sharing in cyber community. In this research, we propose a Volumetric ExpertRank algorithm to indentify the real expert in the cyber community automatically. The algorithm is designed based on the articles evaluation information, volumes of articles, and PageRank algorithm. The research results show that our algorithm would be helpful to indentify expert accurately in the cyber community.