Resumo

COCA BERGOLLA, Yuniesky; ROSELL ACOSTA, Leduan Bárbaro  e  VELAZQUEZ RAMIREZ, Arletis. Model of logic agent with facts-based inference. Rev cuba cienc informat [online]. 2017, vol.11, n.2, pp. 29-45. ISSN 2227-1899.

ABSTRACT Logic programming has remained as one of the main paradigms in research in the area of artificial intelligence. Furthermore the case-based inference is very used for creating expert systems, especially in problems about classification or diagnosis, which can easily determine a target or dependent variable and a set of conditionally independent attributes each other, given the objective. In this paper is proposed a logical agent model with inference based on facts, using the declarative potential of logic programming and the advantages of the case-based reasoning. The model was implemented by developing an open source tool for teaching Artificial Intelligence at the University of Informatics Science. The proposed agent determines, based on their beliefs, the weight of each attribute using elements of information theory, also calculates uncertainty following a Naive Bayes approach, and provides useful information for decision-making in the incorporation of new cases to the knowledgebase. It also performs learning by updating the belief values of each value to make new inferences and incorporate new cases to the base. The model was implemented and tested, the performance results achieved were favorable.

Palavras-chave : Case-based reasoning; logic agent; Prolog.

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )