Utilização de mineração de texto para análise de similaridades nos estudos de impacto ambiental das pequenas centrais hidrelétricas do estado de Mato Grosso

Abstract

The scope of this work is to analyze the Environmental Impact Studies of 25 Small Hydroelectric Power Plants in the state of Mato Grosso, collected from the State Department of Environment website. The data were prepared through lemmatization and preprocessing for textual normalization. The Doc2Vec technique was applied to generate numerical vectors representing the documents and capturing semantic similarities, with the aid of Word2Vec, to create word vector representations. The methodology significantly reducted the number of words, facilitating the extraction of relevant information and improving the analysis of environmental impacts. Document similarity was calculated, and Clusters were formed using graph analysis, identifiyng three distinct Clusters. Each Clustershowed geographical diversity and variability in energy generation capacity while maintaining cohesion in terms of the type and function of the enterprise. This approach proved effective for managing and understanding the impacts of SHPs, highlighting patterns and semantic relationships in the EIS.