Regresar

Linguistic Data Summarization with Multilingual Approach

Abstract:

The development of information systems increases the volume of data and the need for its processing for decision-making. Algorithms are required that allow the discovery of behavior patterns and their interpretability. In this context, the linguistic summarization of data arises as one of the descriptive knowledge discovery techniques with a promising approach to produce summaries from a database using natural language, where authors such as Yager and Zadeh were pioneers and set guidelines in the development of these techniques. In this work, new algorithms are proposed for the generation of linguistic summaries from data that combine different soft computing techniques such as: rough sets, the learning of probabilistic graphs with controlled natural languages for the generation of linguistic summaries with an approach multilingual. In particular, controlled natural languages are proposed for the Spanish, English, Japanese and Arabic languages. The proposed algorithms are compared with other techniques reported in the bibliography and are subject to expert evaluation.