by Alejandra Segura, Claudia Martinez, Christian Vidal, Salvador Sánchez
Abstract:
The information retrieval process use knowledge models for query expansion, recommendation, indexing and/or digital resource labelling. Even though it is generally assumed that the model contains the concept being sought, occasionally the concept is absent from the model (matching problem). In this case, those concepts in the knowledge model that are semantically or syntactically closest to the query may be retrieved instead, thus allowing access to the knowledge represented in the model, so it can be used in an information retrieval process. This paper proposes an algorithm, called the Best Candidate Algorithm (BC Algorithm), to find the closest concept to a query in the matching problem context. In this paper, there were used formal ontologies as models of knowledge. When a query is absent from the model, the algorithm proposes a list of candidates that are sorted based on syntactic and semantic indexes previously defined. The proposal was evaluated through two experiments that led to the conclusion that it is possible to find a closest concept to a query in different domains of knowledge and when there is little information about the query context, specifically, we only know the query and the domain of knowledge where it is immersed.
Reference:
Algorithm to Find the Closest Concept in a Knowledge Model to a Query: Solving the Matching Problem (Alejandra Segura, Claudia Martinez, Christian Vidal, Salvador Sánchez), In International Review on Computers and Software (IRECOS), volume 9, 2014.
Bibtex Entry:
Title = {Algorithm to Find the Closest Concept in a Knowledge Model to a Query: Solving the Matching Problem},
Author = {Alejandra Segura and Claudia Martinez and Christian Vidal and Salvador Sánchez},
Journal = {International Review on Computers and Software (IRECOS)},
Year = {2014},
Number = {10},
Volume = {9},
Abstract = {The information retrieval process use knowledge models for query expansion, recommendation, indexing and/or digital resource labelling. Even though it is generally assumed that the model contains the concept being sought, occasionally the concept is absent from the model (matching problem). In this case, those concepts in the knowledge model that are semantically or syntactically closest to the query may be retrieved instead, thus allowing access to the knowledge represented in the model, so it can be used in an information retrieval process. This paper proposes an algorithm, called the Best Candidate Algorithm (BC Algorithm), to find the closest concept to a query in the matching problem context. In this paper, there were used formal ontologies as models of knowledge. When a query is absent from the model, the algorithm proposes a list of candidates that are sorted based on syntactic and semantic indexes previously defined. The proposal was evaluated through two experiments that led to the conclusion that it is possible to find a closest concept to a query in different domains of knowledge and when there is little information about the query context, specifically, we only know the query and the domain of knowledge where it is immersed.},
ISSN = {1828-6011},
Keywords = {Matching Problem; Information Retrieval; Ontologies},
Url = {http://www.praiseworthyprize.org/jsm/index.php?journal=irecos&page=article&op=view&path%5B%5D=16459}
}