Download Advances in Information Retrieval: 31th European Conference by W. Bruce Croft (auth.), Mohand Boughanem, Catherine Berrut, PDF

By W. Bruce Croft (auth.), Mohand Boughanem, Catherine Berrut, Josiane Mothe, Chantal Soule-Dupuy (eds.)

This ebook constitutes the refereed lawsuits of the thirtieth annual ecu convention on info Retrieval study, ECIR 2009, held in Toulouse, France in April 2009.

The forty two revised complete papers and 18 revised brief papers provided including the abstracts of three invited lectures and 25 poster papers have been conscientiously reviewed and chosen from 188 submissions. The papers are equipped in topical sections on retrieval version, collaborative IR / filtering, studying, multimedia - metadata, professional seek - advertisements, evaluate, opinion detection, net IR, illustration, clustering / categorization in addition to dispensed IR.

Show description

Read or Download Advances in Information Retrieval: 31th European Conference on IR Research, ECIR 2009, Toulouse, France, April 6-9, 2009. Proceedings PDF

Similar european books

Influence and Interests in the European Union: The New Politics of Persuasion and Advocacy

Truly discusses the effect and makes use of of curiosity illustration within the improvement of the ecu method. * Examines the complexities of illustration at ecu point, an essential factor for strength lobbyists and curiosity teams* Charts new developments and concerns comparable to growth, Europeanization and crucial and jap Europe* Contributions via said specialists with a confirmed song checklist of study and e-book during this box, together with seven present and earlier practitioners of european politics with event as lobbyists from both institutional, NGO or company views* areas curiosity illustration in its ancient and theoretical context.

Additional info for Advances in Information Retrieval: 31th European Conference on IR Research, ECIR 2009, Toulouse, France, April 6-9, 2009. Proceedings

Example text

Table 2 (a) illustrates the relationship between the optimal values of b and ncall on the four collections. Table 2 (a) demonstrates that when n decreases, the optimal value of b tends to increase. This demonstrates how the risk adjustment Risk-Aware Information Retrieval 25 Table 3. Performance comparison on six metrics. Three lines in each cell are performance of a language model and our risk-aware approach, and the percentage of gain of our approach over the language model, respectively. Positive and statistically significant improvements are in bold, and in bold and marked with “*”, respectively.

In this case, the relevance return of the list mimics that of the two documents. As a result, the list contains the same amount of uncertainty (risk) as those of the two documents. In this case, risk is not reduced. 4 Document Ranking – A Practical Solution Directly optimizing the objective function in Eq. (6) is computationally expensive. In this section, we present an efficient document ranking algorithm by sequentially optimizing the objective function. It is based on the observation that the larger the rank of a relevant document, the less likely it would be seen or visited by the user.

Given a topic model T M , any document D and word w, w ∈ V, we first calculate a topic model based document model pT M (w|D) by: pT M (w|ti )pT M (ti |D), pT M (w|D) = (5) ti ∈T where pT M (w|ti ) is the multinomial distribution in topic ti , pT M (ti |D) is the probability of observing topic ti in D, and T represents the topic set utilized to calculate this document model. T can either contain all the topics in model T M or just one topic tbest that a document D belongs to with the highest probability: tbest = arg maxti pT M (ti |D).

Download PDF sample

Rated 4.34 of 5 – based on 21 votes