中文論文審查者推薦方法之研究

Research on Recommendation of Chinese Paper Reviewers

羅欣瑜(Hsin-Yu Lo)

指導教授 ： 魏世杰

淡江大學/管理學院/資訊管理學系碩士班/碩士(2007年)

https://doi.org/10.6846/TKU.2007.01002

摘要

隨著傳統期刊逐漸轉變為電子期刊出版模式，除了期刊編輯出版電子化之外，也帶動「線上投稿暨審查系統」之潮流。以OpenConf線上投稿暨審查系統為例，本研究探討其自動推薦論文審查者功能，發現其自動推薦是依據審稿者及投稿文章作者自行勾選的主題群做推薦，侷限於期刊本身自訂的主題範圍。為了打破此侷限，本文方法均以自動粹取關鍵詞來進行，並嘗試用4種不同文章表現法，搭配7種評分方法，包括向量空間模式下的4種相似度計算方法，及OpenConf系統中的3種自動推薦審查者功能，交叉組合出4x7=28個實驗。結果顯示，以向量空間模式的餘弦相似度計算，搭配以摘要文為章表現法的推薦方法，其推薦效果為最好。

關鍵字

向量空間模式 ； 文章表現法 ； 自動推薦 ； 論文審查者 ； OpenConf推薦模式

並列摘要

As more e-journals appear and the e-review process becomes more popular, the demand for automatic recommendation of a good peer reviewer has been ever increasing. With its free and open source policy, OpenConf stands out as a popular system for automatic recommendation of reviewers and serves as a good benchmark for comparison of similar systems. After analysis of its source code, we found that OpenConf represents each reviewer or paper by a set of manually selected topics which are predefined. It also provides three ad hoc scoring methods for matching reviewers and papers. To automate the paper representation, this work evaluates 4 kinds of paper representation which include full text, abstract, title, and author defined keywords. To match reviewers and papers, this work evaluates 7 kinds of scoring methods which include 3 ad hoc methods from OpenConf and 4 variant methods from the traditional vector space model. The result of the 4x7=28 experiments shows that recommendation methods based on the vector space model are better than the 3 ad hoc methods of OpenConf in most document representations. Among them, the paper representation by the abstract combined with the matching by the cosine measure has the highest average precision.

並列關鍵字

Vector Space Model ； Document Representation ； Automatic Recommendation ； Paper Reviewer ； OpenConf Recommendation Model

參考文獻

[3]劉群，李素建，基於《知網》的辭彙語義相似度計算，中文計算語言學期刊，第 7 卷第 2 期, 頁 59-76，2002

[2]邱炯友，李怡萍，學術電子期刊編輯整合平台市場與個案：以教育資料與圖書館學季刊為例，教育資料與圖書館學，第 43 卷第 3 期．頁 327 – 345，2006。

[6]Allan, J., Topic Detection and Tracking: Event-based Information Organization, Kluwer Academic Publishers, 2002.

[8]Chakrabarti, S., Mining the Web: Discovering Knowledge from Hypertext Data, Morgan Kaufmann, 2003.

[9]Kondrak, G., Marcu, D. and Knight, K., “Cognates Can Improve Statistical Translation Models,” Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics, pp. 46-48, 2003.