USE OF TEXT MINING IN INFORMATIONAL SOURCES FOR GROUPS OF RURAL PRODUCERS

Abstract

Observing the need for technological tools to analyze elements of informational fields and recognize links between data sources accessed by certain subjects, this research examines data published for groups of rural producers. It focuses on communications for these groups and government datasets from Brazil. The objective is to verify lexical adherence between units extracted from news for producers in Brazil and government datasets, analyzing informational needs. The methods include a literature review, the 5W1H technique, text mining techniques, and similarity calculation in R. The results indicate insufficient communication, with proximity only in the Credit category. The similarity between the two sources is favorable at an initial level, but there is a need for better contextualization of some words. Future studies aim to contrast this procedure with other similarity measures and apply it to different sources and socioeconomic contexts.