Citation Context

... Section 4) that can obtained by selecting ideal terms (or rejecting unnecessary terms) from long TREC description queries. This potential for improvement motivated past work like Bendersky and Croft =-=[1]-=- and Lease et al. [18] on automatic techniques for query reduction. The former’s approach involved learning to identify key concepts in long queries using a variety of features while the latter focuse...

Citation Context

...en too little to give the algorithm reliable signals about important words. Our work observes that tweets are approximately similar to search queries. It transforms tweets via part-of-speech analysis =-=[2]-=- and inverse document frequency (IDF) metrics into versions more amenable to search engines; it then uses the returned pages to bootstrap additional knowledge about the tweet. TweeTopic can complete t...

Citation Context

...g query term dependencies and using non-uniform query term weighting (beyond idf) can significantly improve retrieval effectiveness, especially on very large collections and for long, complex queries =-=[3, 17, 20]-=-. To our knowledge, no work exists on simultaneously modeling query term dependencies and weighting generic query term concepts (e.g., unigrams, bigrams, etc.) in a unified, trainable framework. It is...

Citation Context

...er because long queries contain many redundant words, the elimination of which is beneficial. This is consistent with current research on long queries which states that either downweighting bad terms =-=[4, 17]-=- or dropping them [15] tends to help with long queries. 6. RELATED WORK Query reformulation has a very long history, dating back to the earliest relevance feedback techniques [22] where queries are mo...

Citation Context

...ting retrieval methods, in general, perform worse for the long queries than for the short ones; a result which has been consistently and independently affirmed both in the experiments on TREC corpora =-=[4, 18]-=- and in the web search setting [9]. This result has been attributed to the rarity of the long queries [9], a lack of sufficient natural language parsing by the underlying retrieval method [19], term r...

Citation Context

...ing. However, in comparison to shorter and less informative keyword queries like “World Bank Criticism”, verbose queries actually tend to yield worse retrieval accuracy with current retrieval methods =-=[2, 8]-=-. At the Reliable Information Access workshop [4], detailed failure analysis of2 Matthew Lease, James Allan, and Bruce Croft verbose queries found that in most cases, significantly improved retrieval...

Citation Context

... beyond document frequency on the importance of a query concept. Flexible query term weighting is particularly important for verbose queries that often contain a mix of key and complementary concepts =-=[5, 14]-=-. Second, most of the importance weighting research has been applied to single terms (i.e., unigrams). Relatively little research has been done to investigate the appropriateness of IDF and other unsu...

Citation Context

... word queries dropped 3% [1]. Unfortunately, web search engines perform poorly on longer queries when compared to shorter ones [4]. Several past works have focused on improving long query performance =-=[3, 15, 14, 16, 13, 7]-=- (see Section 2). They can be broadly classified into query re-weighting and query reduction approaches. In query re-weighting, the original query terms are assigned different weights before the query...

Citation Context

...7 topical categories based on a list of manually pre-categorized queries; Jansen et al. [46] use commonly occurring multimedia terms to categorize audio, video, and image queries; and Huurnink et al. [47] utilize structured data from clicked results to link queries in a multimedia archive to an in-house thesaurus. Many applications of (semantic) query analysis have been proposed, such as disambiguation [48,49] and rewriting. Jansen et al. [11] use query logs to determine which queries or query rewrites occur frequently. Others perform query analysis and try to identify the most relevant terms [50], to predict the query’s performance a priori [51], or combine the two [52]. Bendersky and Croft [50] use part-of-speech tagging and a supervised machine learning technique to identify the ‘‘key noun phrases’’ in natural language queries. Key noun phrases are phrases that convey the most information in a query and contribute most to the resulting retrieval performance. Our approach differs in that we link queries to a structured knowledge base instead. We incorporate and evaluate several of the features proposed in [50–52] on our task below.3. The task The query mapping task that we address in...

Citation Context

...Information Retrieval 1. INTRODUCTION The use of verbose (or long) queries helps users to express their information need naturally and saves efforts in choosing keywords. Unfortunately, previous work =-=[1, 11]-=- has shown that current search engines cannot handle verbose queries well. It seems that the additional information provided in Permission to make digital or hard copies of all or part of this work fo...