Query Expansion based on Concept Clique for Markov Network Information Retrieval Model

Abstract

Query expansion is a common technique used to improve retrieval effectiveness. In this paper, we propose a novel query expansion technique based on concept clique for Markov network information retrieval model. This technique strengthens the simple relationships between terms in the following two ways:(1) terms in a clique express a similar concept and will be added into query expansion, so that it is effective to expanded into some terms with low similar to query terms but highly related to query topic; (2) query term dependencies are used to select concept cliques as candidates. The selection of concept cliques based on a connected graph is effective to avoid topic drift during expanding polysemous query terms. Experiments on several collections show that new approach makes significant improvements and more effective on collections with polysemous terms.

Método para autocompletar consultas basado en cadenas de Markov y la ley de Zipf

“…Algunos métodos utilizan cadenas de Markov para seleccionar dichos términos y las probabilidades de transición (es decir, las probabilidades de pasar de una palabra a otra) se estiman, en [15] mediante los tipos de relación semántica (polisemia, antonimia, sinonimia, etc.) entre las palabras, en [16] con la combinación de múltiples fuentes de conocimiento sobre sus relaciones y en [17] mediante las relaciones entre los significados de las palabras, todos obtienen excelentes resultados.…”

An Information Retrieval Expansion Model Based on Quasi-Clique

“…And then for the second selection, we will choose -Quasi-Clique which sets q i as a core node. When the value of  is set to 1, a 1-Quasi-Clique graph is a complete graph such as clique which is adopted in the model [7]. As shown in the Fig.1 above, if the value of  is set to 0.5, then the list of QuasiCliques containing t 1 is {(t 1 , t 2 , t 3 , t 5 ), (t 1 , t 4 , t 6 , t 7 )}.Therefore, we can see that the query term t 1 will get more relevant candidates such as terms t 2 and t 6 .Similarly, every query term will get its candidate list of Quasi-Cliques for query expansion.…”

“…The models proposed in [6] perform outstanding and make significant improvements. A novel query expansion technique based on concept clique for Markov network information retrieval model was proposed in [7], which integrates query term dependency to the selection of concept clique. There is a hypothesis in [7] that if query terms within a query have relationships, their concept cliques as candidates will be in a connected graph.…”

“…A novel query expansion technique based on concept clique for Markov network information retrieval model was proposed in [7], which integrates query term dependency to the selection of concept clique. There is a hypothesis in [7] that if query terms within a query have relationships, their concept cliques as candidates will be in a connected graph. Therefore, when expanding dependent query term, although some single concept cliques with high similarity to original query terms, they will be considered as noise and be not taken into account.…”

Mobile Search Engine using Clustering and Query Expansion

“…The goal of query expansion is to improve retrieval effectiveness, it does this by finding terms that the relevant web pages are indexed by and adding them to the original query [22] [23]. By adding additional search terms we get a more specific query that is less ambiguous and will return web pages that reflects the underlying information needed.…”