手法を比較

	Relevance Feedback Evaluation	Query Expansion Evaluation
分野	Library Information Science	Library Information Science
系統	Process / pipeline	Process / pipeline
提唱年	1990	2008
提唱者	Gerard Salton & Chris Buckley (building on J. J. Rocchio)	Information retrieval evaluation tradition (Salton; Manning, Raghavan & Schütze)
種類	Evaluation pipeline for relevance-feedback query reformulation	Evaluation pipeline for query-expansion effectiveness
原典	Salton, G., & Buckley, C. (1990). Improving retrieval performance by relevance feedback. Journal of the American Society for Information Science, 41(4), 288-297.	Manning, C. D., Raghavan, P., & Schütze, H. (2008). Introduction to Information Retrieval. Cambridge University Press. ISBN: 9780521865715
別名	Rocchio Feedback Evaluation, Feedback Effectiveness Measurement, Residual Collection Evaluation, Relevance Feedback Assessment	Query Reformulation Evaluation, Term Expansion Assessment, Expansion Effectiveness Measurement, Automatic Query Expansion Evaluation
関連	3	3
概要	Relevance feedback evaluation measures how much a retrieval system improves when it reformulates a query using user judgments on the first results. The technique that defined the field is Rocchio's vector-space feedback, in which documents the user marks relevant pull the query vector toward themselves and documents marked non-relevant push it away; Salton and Buckley's 1990 study systematized its evaluation and showed substantial effectiveness gains. The central methodological challenge is fairness: because the user has already seen and judged some documents, naively re-scoring the whole collection rewards the system for re-finding documents it was just told about. Residual-collection and frozen-rank evaluation solve this by measuring improvement only on documents the user has not yet seen.	Query expansion evaluation measures whether adding terms to a user's query — drawn from a thesaurus, from corpus co-occurrence statistics, or from relevance or pseudo-relevance feedback — actually improves retrieval. Expansion attacks the vocabulary-mismatch problem, where relevant documents use words the searcher did not, and it tends to raise recall by bringing in synonymous and related terms. But it can also lower precision by introducing ambiguity, and it can help some queries while badly hurting others. Sound evaluation therefore reports not just the average effectiveness change but the recall-precision trade-off and a robustness analysis of how many individual queries were helped versus harmed.