Method combination for document filtering

Hull, David A.; Pedersen, Jan O.; Schütze, Hinrich

doi:10.1145/243199.243275

Cited by 87 publications

(56 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…As combination methods work best when the results being combined are generated independently (Saracevic and Kantor 1988, Turtle and Croft 1991, Hull et al 1996, Dietterich 1997, it is conceivable that the improvements over individual methods could be higher if we weakened some experimental factors that are likely to increase the correlation between the term-relevance estimates of the different functions, for instance by varying the number of documents used for query expansion or their representations.…”

Section: Discussionmentioning

confidence: 99%

Improving retrieval feedback with multiple term-ranking function combination

Carpineto

Romano

Giannini

2002

ACM Trans. Inf. Syst.

View full text Add to dashboard Cite

and VITTORIO GIANNINI WIND Telecomunicazioni S.p.A ________________________________________________________________________In this paper we consider methods for automatic query expansion from top retrieved documents (i.e., retrieval feedback) which make use of various functions for scoring expansion terms within Rocchio's classical reweighting scheme. An analytical comparison shows that the retrieval performance of methods based on distinct term-scoring functions is comparable on the whole query set but considerably differs on single queries, consistent with the fact that the ordered sets of expansion terms suggested for each query by the different functions are largely uncorrelated. Motivated by these findings, we argue that the results of multiple functions can be merged, by analogy with ensembling classifiers, and present a simple combination technique based on the rank values of the suggested terms. The combined retrieval feedback method is effective not only with respect to unexpanded queries but also to any individual method, with notable improvements on the system's precision. Furthermore, the combined method is robust with respect to variation of experimental parameters and it is beneficial even when the same information needs are expressed with shorter queries.

show abstract

Section: Discussionmentioning

confidence: 99%

Improving retrieval feedback with multiple term-ranking function combination

Carpineto

Romano

Giannini

2002

ACM Trans. Inf. Syst.

View full text Add to dashboard Cite

show abstract

Section: Fix the Fuzzy Area Between Categories By The Multi-class Baymentioning

confidence: 99%

“…Let: We try to transfer the multi-class TC described by (10) into a two-class TC described by (14). Formula (14) means that the binary-valued multi-class Naïve Bayesian Classifier can be approximately regarded as searching a separate line in a twodimensional space with MV max_F being the X-coordinate and MV max_S being the Ycoordinate. The distance from a given document, represented as a point (x, y) with the values of x and y calculated according to (12) and (13) respectively, to the separate line in this two-dimensional space will be:…”

Section: Fix the Fuzzy Area Between Categories By The Multi-class Baymentioning

confidence: 99%

“…For example, Yang et al [13] used simple equal weights for normalized score of each classifier output so as to integrate multiple classifiers linearly in the domain of Topic Detection and Tracking; Hull at al. [14] used linear combination for probabilities or log odds scores of multiple classifier output in the context of document filtering. Larkey et al [15] used weighted linear combination for system ranks and scores of multiple classifier output in the medical document domain; Li and Jain [16] used voting and classifier selection technique including dynamic classifier selection and adaptive classifier.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Classifying Chinese Texts in Two Steps

Fan

Sun

Choi

et al. 2005

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. This paper proposes a two-step method for Chinese text categorization (TC). In the first step, a Naïve Bayesian classifier is used to fix the fuzzy area between two categories, and, in the second step, the classifier with more subtle and powerful features is used to deal with documents in the fuzzy area, which are thought of being unreliable in the first step. The preliminary experiment validated the soundness of this method. Then, the method is extended from two-class TC to multi-class TC. In this two-step framework, we try to further improve the classifier by taking the dependences among features into consideration in the second step, resulting in a Causality Naïve Bayesian Classifier.

show abstract

“…Although our focus in this paper is primarily on combination techniques for improving retrieval effectiveness, combination has been applied to a number of related tasks, such as filtering (Hull et al, 1996) and categorization (Lewis and Hayes, 1994;Larkey and Croft, 1996), and has been studied in other fields such as machine learning (Mitchell, 1997). We will refer to work in these areas in a number of sections of the paper.…”

Section: Introductionmentioning

confidence: 99%

Combining Approaches to Information Retrieval

Croft

The Information Retrieval Series

113

135

View full text Add to dashboard Cite

The combination of different text representations and search strategies has become a standard technique for improving the effectiveness of information retrieval. Combination, for example, has been studied extensively in the TREC evaluations and is the basis of the "meta-search" engines used on the Web. This paper examines the development of this technique, including both experimental results and the retrieval models that have been proposed as formal frameworks for combination. We show that combining approaches for information retrieval can be modeled as combining the outputs of multiple classifiers based on one or more representations, and that this simple model can provide explanations for many of the experimental results. We also show that this view of combination is very similar to the inference net model, and that a new approach to retrieval based on language models supports combination and can be integrated with the inference net model.

show abstract

Method combination for document filtering

Cited by 87 publications

References 26 publications

Improving retrieval feedback with multiple term-ranking function combination

Improving retrieval feedback with multiple term-ranking function combination

Classifying Chinese Texts in Two Steps

Combining Approaches to Information Retrieval

Contact Info

Product

Resources

About