應用負相關回饋資訊於文件重排序之分析

An analysis of the application of non-relevance feedback in document ranking

張弘杰(Hong-Kiat Chong)

指導教授 ： 周世傑

國立中央大學/管理學院/資訊管理學系/碩士(2012年)

摘要

負相關回饋的資訊雖然被認為可利用價值不高, 但它在促進資訊擷取的效能上，仍然有可利用之處。有研究嘗試利用負相關回饋資訊於文件檢索結果的重排序，並且初步顯現效能，本研究依據理論探討，認為負相關回饋資訊於文件檢索結果重排序的應用，可能受資料離散性等資料分佈情境的影響, 因此，分析資料離散等資料分佈情境與負相關回饋資訊的應用成為本研究的目的。為此，本研究針對初始檢索結果進行資料離散等分析, 確定資料分佈情境與負相關回饋資訊的應用是否有直接關聯, 提出了資料分佈情境對負相關回饋資訊應用的影響。實驗結果指出, 文件資料的離散性並沒有與負相關回饋資訊應用的效能有線性關係, 但是相關與不相關文件之間差異性小的文件型態會對負相關回饋資訊的應用有不良影響。根據這種情況, 本研究提出了數個未來研究發展的方向。

關鍵字

負相關回饋 ； 資訊檢索 ； 文件重排序 ； 文件分析

並列摘要

Although the information of non-relevance feedback information is thought as not much useful in information retrieval, it still can be applied. Some research tried using non-relevance feedback information in document re-ranking. In this research, our goal is to disclose the relation between the data distribution and the application of non-relevance feedback according to the theory that we had studied. In order to do so, we focus on the analysis of the distribution of initial retrieval result, and the direct links between distribution scenario and the application of non-relevance feedback. The final result shows that the distribution of the text data and the application of non-relevance feedback doesn’t exist linear relationship and the significance of difference between relevance and non-relevance in dataset could affect the application of non-relevance feedback. Base on this result, our research propose some direction in future study.

並列關鍵字

Document re-ranking ； Information retrieval ； Non-relevance feedback ； Document Analysis

參考文獻

[1] Brin, S., Page, L. (1988). The anatomy of a large-scale hypertextual Web search engine, Computer Networks and ISDN Systems, Volume 30, Issues 1–7.

[3] Chou, S., & Chang, W. (2008). CyberIR – A Technological Approach to Fight Cybercrime. Lecture Notes in Computer Science, 5075, 32–43.

[6] Yang, C.C.; Dorbin Ng, T. (2011). Analyzing and Visualizing Web Opinion Development and Social Interactions with Density-Based Clustering. Systems, Man and Cybernetics, Part A: Systems and Humans, IEEE Transactions on, vol.41, no.6, pp.1144 - 1155.

[7] Liddy, E. D. (1998). Enhanced Text Retrieval Using Natural Language Processing. Bulletin of the American Society for Information Science, 24(4), 14–16.

[9] Can, F. and Ozkarahan, E. A. (1990). Concepts and effectiveness of the cover-coefficient-based clustering methodology for text databases. ACM Trans. Database Syst. 15, 483-517.