Regresar

Software ecosystem for learning and decision making based on linguistic summarization of data

Abstract:

Introduction: The increase in data volume in dissimilar decision-making scenarios increases the need for techniques for the discovery of non-trivial dependencies hidden in the data. Linguistic data summaries are identified as one of the branches of artificial intelligence that allow generating linguistic summaries with application in different areas. Objective: To develop an ecosystem for learning and decision-making, based on linguistic data summary techniques under a multilingual approach, improving the efficiency of existing systems. Methods: It is applied the logical historical method that allows the identification of the main opportunities for improvement of existing methods. They are proposed new algorithms and methods for the construction of linguistic summaries. Finally, they are applied empirical methods, and the results are validated by combining data triangulation techniques and methods. Results: As a result of the research, four new algorithms were obtained that improved the efficiency and effectiveness of other algorithms reported in the literature. The proposed algorithms stand out for their facilities for multilingual work, the use of neutrosophic techniques, and the treatment of correlated variables. Conclusions: From the comparisons it is concluded that the RST_LDS algorithm demonstrates the complementarity and importance of combining different techniques in the discovery of linguistic summaries. The new indicators for the evaluation of linguistic summaries improve the treatment of indeterminacy and falsity by complementing the indicators reported in the bibliography.