URI permanente para esta colección:

Artículos en la categoría Fuzzy publicados en las Actas de las XXII Jornadas sobre Programación y Lenguajes (PROLE 2023).
Notificar un error en esta colección


Envíos recientes

Mostrando 1 - 3 de 3
  • Artículo
    On the Safe Execution of Symbolic Similarity-based Fuzzy Logic Programs
    Moreno, Gines; Riaza Valverde, José Antonio. Actas de las XXII Jornadas sobre Programación y Lenguajes (PROLE 2023), 2023-09-12.
    Introducing “Fuzzy Aggregators and Similarity Into a Logic Language” leads to FASILL, whose symbolic extension (called sFASILL) is able to manage both concrete and unknown (symbolic constants) truth degrees, similarity annotations and fuzzy connectives. The symbolic execution of sFASILL programs helps to guess the impact of such unknown symbols on the outputs before instantiating them with concrete values in program rules and similarity relations. In this paper we reason about the constraints that symbolic substitutions must satisfy in order to preserve the set of fuzzy computed answer obtained at execution time before or after instantiating a given sFASILL program. Based on such safeness conditions, we have formally proved two theoretical results with different levels of generality/practicability for correctly executing symbolic sFASILL and instantiated FASILL programs.
  • Artículo
    Fuzzy Retrieval of Linked Open Data
    Almendros-Jimenez, Jesus M.; Becerra-Teron, Antonio; Moreno, Gines. Actas de las XXII Jornadas sobre Programación y Lenguajes (PROLE 2023), 2023-09-12.
    The technologies associated with the Semantic Web facilitate the publication of Open Data as well as their integration into the Linked Open Data cloud for which RDF format has been chosen. The promoted W3C RDF query language SPARQL became very popular and has been subject of study in recent years, with the goal of improving SPARQL engines in terms of functionality and usability. In this line of research, our research group has developed a fuzzy extension of SPARQL, called FSA-SPARQL (Fuzzy Sets and Aggregators based SPARQL). Here we go a step forward, adapting FSA-SPARQL for querying Linked Open Data with a fuzzy taste. Three goals are pursued, (i) fuzzification of Linked Open Data datasets, (ii) to extend FSA-SPARQL fuzzy aggregation and (iii) to equip FSA-SPARQL with quantification. Fuzzification, that is, interpretation of Linked Open Data datasets as fuzzy sets is made automatically at querying time, using trapezoidal membership functions. Powerful fuzzy aggregation and quantification operators are introduced in FSA-SPARQL based on fuzzy cardinals of fuzzy sets. As a result of the goals, FSA-SPARQL can be used for the retrieval of Linked Open Data datasets in real time, combining fuzzy aggregation and quantification. The proposed extension for FSA-SPARQL has been implemented and the system is available at our Web site.
  • Resumen
    Bousi∼Prolog: Design and implementation of a proximity-based fuzzy logic programming language.
    Julián-Iranzo, Pascual; Saenz-Perez, Fernando. Actas de las XXII Jornadas sobre Programación y Lenguajes (PROLE 2023), 2023-09-12.
    The fuzzy logic programming language Bousi∼Prolog extends Prolog with a weak unification algorithm based on proximity relations and truth degree annotations. The weak unification algorithm makes the search for answers more flexible, while rule annotations make possible knowledge-based applications where the rules may be uncertain. In this paper, after recalling the main concepts supporting this language, we detail its design and implementation. We describe the implementation of its operational semantics, which is based on compiling programs and queries into Prolog, and those important features that makes it more applicable: fuzzy sets, integration with WordNet and efficiency techniques. The result is a high-level open-source implementation of the Bousi∼Prolog system, written on top of SWI-Prolog, and publicly available. We also summarise some experiments measuring its performance compared to other systems. [Work already published: Expert Syst. Appl. 213(Part): 118858 (2023) ]