Link-based and content-based evidential information in a belief network model

Abstract

This work presents an information retrieval model developed to deal with hyperlinked environments. The model is based on belief networks and provides a framework for combining information extracted from the content of the documents with information derived from cross-references among the documents. The information extracted from the content of the documents is based on statistics regarding the keywords in the collection and is one of the basis for traditional information retrieval (IR) ranking algorithms. The information derived from crossreferences among the documents is based on link references in a hyperlinked environment and has received increased attention lately due to the success of the Web. We discuss a set of strategies for combining these two types of sources of evidential information and experiment with them using a reference collection extracted from the Web. The results show that this type of combination can improve the retrieval performance without requiring any extra information from the users at query time. In our experiments, the improvements reach up to 59% in terms of average precision figures.

A Model for Information Retrieval Based on Possibilistic Networks

“…This system evaluates the belief in a document with respect to a query, and a list of weighted documents is retrieved. Belief Networks [11] [17] have been used to "model knowledge derived from past queries and combine it with the vector space model" [11]. The ranking of a document is based on the similarity between document d j and query Q, computing the probability P (d j = 1/Q = 1).…”

“…For an efficient Information Retrieval System (IRS) these two sets must be equal as often as possible. The relevance of a document to a query is usually interpreted by most of IR models, vector space [14], probabilistic [12][13] [18], inference and belief networks [20][11] [17], as a score computed by summing the inner products of term weights in the documents and query representations. Whatever the used model, the response to a user need is a list of documents ranked according to a relevance value.…”

A Layered Bayesian Network Model for Document Retrieval

“…The Ribeiro and Reis' model [7,6,11] is designed to simulate the Vector Space, Boolean and Probabilistic models. Their network is composed of three types of nodes: document nodes, concept nodes, and the query node.…”

A query processing algorithm for hierarchical Markov networks

“…Since the hierarchical Markov network has many advantages over the Markov network representation [15], and given the explanation of the favorable experimental results at the end of Section 4, we feel that the hierarchical Markov network representation could eventually become the standard representation of Bayesian networks. Thus, the encouraging results reported in this paper are useful to any work applying BNs, including traditional information retrieval [2,7,10,16], web search [9], user profiling [11], multi-agents [5,12,17] and e-commerce [4].…”

“…In particular, Bayesian networks (BNs) [6] use the notion of probabilistic conditional independence [13] to facilitate the acquisition of a joint probability distribution. Several researchers have naturally suggested that Bayesian networks be applied in traditional information retrieval [2,7,10,16], web search [9], user profiling [11], multi-agents [5,12,17] and e-commerce [4].…”