Download A Generative Theory of Relevance by Victor Lavrenko PDF

By Victor Lavrenko

A smooth details retrieval procedure should have the potential to discover, arrange and current very varied manifestations of knowledge – resembling textual content, images, movies or database files – any of that could be of relevance to the person. although, the idea that of relevance, whereas doubtless intuitive, is absolutely tough to outline, and it really is even tougher to version in a proper way.

Lavrenko doesn't try to bring about a brand new definition of relevance, nor offer arguments as to why any specific definition should be theoretically better or extra whole. as a substitute, he is taking a generally approved, albeit a little conservative definition, makes a number of assumptions, and from them develops a brand new probabilistic version that explicitly captures that concept of relevance. With this ebook, he makes significant contributions to the sector of data retrieval: first, a brand new option to examine topical relevance, complementing the 2 dominant types, i.e., the classical probabilistic version and the language modeling method, and which explicitly combines records, queries, and relevance in one formalism; moment, a brand new strategy for modeling exchangeable sequences of discrete random variables which doesn't make any structural assumptions in regards to the info and which could additionally deal with infrequent events.

Thus his publication is of significant curiosity to researchers and graduate scholars in info retrieval who specialise in relevance modeling, score algorithms, and language modeling.

Show description

Read Online or Download A Generative Theory of Relevance PDF

Similar structured design books

The .NET Developer's Guide to Directory Services Programming

Lively listing is a vital delivering through Microsoft, essentially to be used inside its . internet Framework. What Kaplan and Dunn recommend here's that the programmer-level documentation for lively listing being provided by means of Microsoft is just a little awkward to exploit and comprehend. So this e-book is on the market. The context is tips on how to code LDAP within the namespace of procedure.

Primality Testing in Polynomial Time: From Randomized Algorithms to "PRIMES Is in P"

On August 6, 2002,a paper with the name “PRIMES is in P”, by way of M. Agrawal, N. Kayal, and N. Saxena, seemed at the web site of the Indian Institute of expertise at Kanpur, India. during this paper it used to be proven that the “primality problem”hasa“deterministic set of rules” that runs in “polynomial time”. checking out even if a given quantity n is a primary or now not is an issue that was once formulated in precedent days, and has stuck the curiosity of mathema- ciansagainandagainfor centuries.

Automata, Languages and Programming: 36th International Colloquium, ICALP 2009, Rhodes, Greece, July 5-12, 2009, Proceedings, Part II

The two-volume set LNCS 5555 and LNCS 5556 constitutes the refereed lawsuits of the thirty sixth overseas Colloquium on Automata, Languages and Programming, ICALP 2009, held in Rhodes, Greece, in July 2009. The 126 revised complete papers (62 papers for song A, 24 for tune B, and 22 for music C) awarded have been conscientiously reviewed and chosen from a complete of 370 submissions.

Rationale-Based Software Engineering

Many selections are required during the software program improvement strategy. those judgements, and to some degree the decision-making procedure itself, can top be documented because the intent for the procedure, so one can exhibit not just what used to be performed in the course of improvement however the purposes at the back of the alternatives made and possible choices thought of and rejected.

Additional resources for A Generative Theory of Relevance

Sample text

E. the semantic correspondence between a given request and a given document. We will not be addressing issues of presentation, novelty, or suitability to a particular task. 3 Existing Models of Relevance This book is certainly not the first endeavor to treat relevance in probabilistic terms. Some of the more prominent examples are the 2-Poisson indexing model developed independently by Bookstein and Swanson [15, 16] and Harter [52], the probabilistic retrieval model of Robertson and Sparck Jones [117], the probabilistic flavors [123] of Van Rijsbergen’s logical model [139], the inference-network model developed by Turtle and Croft [135, 134], the language modeling approach pioneered by Ponte and Croft [106, 105] and further developed by others [90, 56], and the recent risk minimization framework of Zhai and Lafferty [68, 157].

However, since 1976 the model has been re-formulated a number of times to a degree where it can hardly be called “binary” and, as we shall argue later on, the term “independence” is also questionable. The model is also known as the Okapi model , the City model, or simply as the probabilistic model . A very detailed account of the recent state of the model is provided by the original authors in [131, 132]. What follows is our interpretation of the model. An attentive reader may find that our description is different in two ways from the way the model is presented by the authors [117, 131, 132].

2. Requests. We are concerned only with observable natural-language requests, for which we can obtain relevance judgments. However, our model will involve a notion similar to the real information need (RIN), which will play the role of a latent variable. We will use the words “query” and “request” interchangeably. 3. Non-interactive. We will not be modeling any evolution of user’s information need. Our model explicitly accounts for the fact that a single information need can be expressed in multiple forms, but we do not view these in the context of an interactive search session.

Download PDF sample

Rated 4.35 of 5 – based on 42 votes