相關回饋段落對於文件檢索效能之影響

The Impact of Relevance Feedback Passages on Documents Retrieval Performance

楊朝傑(Chao-Jie Yang)

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(2011年)

摘要

在文件檢索系統當中，透過相關回饋方法與查詢字詞擴張的應用，可使檢索系統的效能獲得一定程度的改善。但所謂的相關文件，其內容並非完全與使用者資訊需求相符，其中還是有若干部分是非相關或是無意義的雜訊區塊，而這些區塊內資訊就會影響查詢擴張後的字詞，使得擴張後的查詢仍無法獲得更佳的文件檢索準確率，因此本研究將運用相關回饋段落資訊來進行查詢擴張，並探討不同的段落挑選方式下對於文件檢索效能的影響。實驗中會先將相關回饋中的文件進行段落切割，並依據不同的段落挑選組合模式，將這些段落做為擴張查詢字詞的資訊來源，最後以第二次的文件檢索的準確率來探討不同段落大小與挑選方式造成的影響。
本研究的實驗結果顯示，藉由相關回饋當中段落資訊的應用，能讓檢索系統進行查詢擴張時排除掉更多非相關字詞，使得二次文件檢索的準確率有更佳的提升，且其改善的後的效果亦較以全文回饋的方式更為有效。

關鍵字

相關回饋 ； 段落挑選 ； 查詢擴張 ； 文件檢索

並列摘要

In information retrieval systems, the system performance can be improved by the application of relevance feedback and query expansion. In fact, the contents of the relevant documents do not always match the information needs of the users. There are still several non-relevant and meaningless noise blocks in the documents and those noises will affect the performance of document retrieval. In this research, it attempts to eliminate the effect of the noises from relevance feedback by using passages information, and also discuss this impact of different passages combination on the performance of second document retrieval. In the experiments, our system first splits the contents of the documents into passages, and then selects passages according to the selection method to use in query expansion. Finally, we use the precision of second document retrieval to discuss the influence of different passage selection methods.
According to the results of experiments, the system performance can be improved by using passages information. The precision of second document retrieval which using passages information is better than the performance which using full-text information. This study has proved that the passages information is very useful on query expansion.

並列關鍵字

query expansion ； document retrieval ； relevance feedback ； passage selection

參考文獻

[7] Kaszkiel, M., & Zobel, J. (1997). Passage retrieval revisited. SIGIR Forum, 31(SI), 178-185.

[9] Krikon, E., Kurland, O., & Bendersky, M. (2010). Utilizing inter-passage and inter-document similarities for reranking search results. ACM Transactions on Information Systems, 29(1), 1-28.

[13] Porter, M. F. (1980). An algorithm for suffix stripping. Program: electronic library and information systems, 14(3), 130-137.

[15] Quiroga, L. M., & Mostafa, J. (2002). An experiment in building profiles in information filtering: the role of context of user relevance feedback. Information Processing & Management, 38(5), 671-694.

[16] Rocchio, J. J. (1971). Relevance feedback in information retrieval. In G. Salton (Ed.), The SMART Retrieval System: Experiments in Automatic Document Processing (pp. 313-323): Prentice-Hall, Englewood Cliffs NJ.