By Victor Lavrenko

ISBN-10: 3540893636

ISBN-13: 9783540893639

ISBN-10: 3540893644

ISBN-13: 9783540893646

A sleek info retrieval process should have the potential to discover, arrange and current very diverse manifestations of knowledge – similar to textual content, photos, video clips or database files – any of that may be of relevance to the person. besides the fact that, the concept that of relevance, whereas possible intuitive, is absolutely tough to outline, and it is even tougher to version in a proper way.

Lavrenko doesn't try to bring on a brand new definition of relevance, nor offer arguments as to why any specific definition should be theoretically improved or extra entire. as an alternative, he's taking a greatly approved, albeit a bit of conservative definition, makes a number of assumptions, and from them develops a brand new probabilistic version that explicitly captures that idea of relevance. With this booklet, he makes significant contributions to the sector of knowledge retrieval: first, a brand new method to examine topical relevance, complementing the 2 dominant versions, i.e., the classical probabilistic version and the language modeling technique, and which explicitly combines files, queries, and relevance in one formalism; moment, a brand new technique for modeling exchangeable sequences of discrete random variables which doesn't make any structural assumptions in regards to the facts and which may additionally deal with infrequent events.

Thus his booklet is of significant curiosity to researchers and graduate scholars in details retrieval who specialise in relevance modeling, rating algorithms, and language modeling.

Show description

Read Online or Download A Generative Theory of Relevance PDF

Best structured design books

Alberto Belussi, Barbara Catania, Eliseo Clementini, Elena's Spatial Data on the Web: Modeling and Management PDF

Spatial information is key in quite a lot of software domain names this day. whereas geographical functions stay the most important goal sector, spatial houses are required in different contexts equivalent to computer-aided layout, robotics and photo processing. linked to those is the always becoming variety of disbursed processing architectures, in accordance with, for instance, grid platforms, sensor information networks, and customized clever units.

Transactions on Computational Systems Biology XII: Special by Rainer Breitling, Corrado Priami, David Gilbert, Monika PDF

The LNCS magazine Transactions on Computational structures Biology is dedicated to inter- and multidisciplinary study within the fields of desktop technology and existence sciences and helps a paradigmatic shift within the ideas from machine and data technological know-how to deal with the hot demanding situations coming up from the platforms orientated perspective of organic phenomena.

Robert Schaefer, Carlos Cotta, Joanna Kolodziej, Günter's Parallel Problem Solving from Nature, PPSN XI: 11th PDF

This publication constitutes the refereed court cases of the eleventh overseas convention on Parallel challenge fixing from Nature - PPSN XI, held in Kraków, Poland, in September 2010. The 131 revised complete papers have been rigorously reviewed and chosen from 232 submissions. The convention covers a variety of subject matters, from evolutionary computation to swarm intelligence, from bio-inspired computing to genuine international purposes.

Download e-book for iPad: Principles of Distributed Systems: 18th International by Marcos K. Aguilera, Leonardo Querzoni, Marc Shapiro

This booklet constitutes the refereed court cases of the 18th foreign convention on ideas of disbursed structures, OPODIS 2014, Cortina d'Ampezzo, Italy, in December 2014. The 32 papers awarded including invited talks have been conscientiously reviewed and chosen from ninety eight submissions. The papers are prepared in topical sections on consistency; disbursed graph algorithms; fault tolerance; types; radio networks; robots; self-stabilization; shared information constructions; shared reminiscence; synchronization and common building.

Extra resources for A Generative Theory of Relevance

Sample text

We are going to argue that the classical Binary Independence Model really does not assume word independence, and consequently that there is no benefit in trying to relax the non-existent assumption. Our explanation is an extension of a very important but almost universally ignored argument made by Cooper in [29]. 2), we only need to assume linked dependence between words Di , and that assumption is substantially weaker than independence. Cooper’s argument is as follows. Consider the case of a two-word vocabulary V={a, b}, and suppose we do not assume independence, so P1 (Da , Db ) is their joint distribution in the relevant class, P0 (Da , Db ) is the same for the non-relevant class.

In order to take account of frequencies researchers have had to assume a different event space. Virtually every publication concerning language modeling in IR [126, 127, 90, 11, 152, 58, 56, 158, 159] presumes the following representation, though it is rarely stated in formal terms. Assume V is a vocabulary of NV distinct words. Both documents and queries are viewed as strings (sequences) over V. A document D of length m is a sequence of m random variables Di , each taking values in the vocabulary V.

If he picks word v as a keyword for document d, then we say that d belongs to the elite class of v. Otherwise d belongs to the non-elite class. 3 Existing Models of Relevance 21 in the elite class of v are likely to contain many repetitions of v, while in the non-elite class v would primarily occur by chance. Harter assumed that frequency of v in both classes follows a Poisson distribution, but that the mean is higher in the elite class. Under this assumption, the frequency of v in the collection as a whole would follow a mixture of two Poissons: P (Dv =dv ) = P (E=1) v v e−μ0,v μd0,v e−μ1,v μd1,v + P (E=0) dv !

Download PDF sample

A Generative Theory of Relevance by Victor Lavrenko


by Charles
4.2

Rated 4.92 of 5 – based on 34 votes