Financial analysis using artificial neural networks applied to the electric energy sector ; Análise financeira com utilização das redes neurais artificiais aplicadas ao setor de energia elétrica

This work collects financial data related to the energy distribution process in Brazil with the aid and use of Artificial Neural Networks (ANNs). Thus, the general objective of the article was to build, train and validate an ANN model, with the specific objective of verifying the correlations between the selected variables with significant security. The software used to analyze the data was SPSS - Statistical Package for the Social Sciences, in order to demonstrate the behavior of the independent variables, the model presented satisfactory values and observations for its validation. The results presented are based on the periods of its operation, containing perspectives, assessments and the importance of managers regarding future situations and financial abilities that may influence the entity's enterprise. The analysis of financial statements has the ability to gather useful information from the financial statements for users in the accounting industry. It involves elements that are capable of modifying the company's actual performance, its sustainability capacity and changes that it may have in the course of its activities. Through the analysis of financial statements, it is possible to measure the company's situation through indicators. Check the degree of liquidity, indebtedness and changes in equity during the years studied, in order to find mechanisms for decision-making by administrators. It is also allowed to compare the entity's evolution or fall rates with those of other companies that operate in the same fields of activity, in order to keep managers informed about the evolution of the market in which they operate. ; Este trabalho refere-se a coleta de dados financeiros relacionados ao processo de distribuição energética no Brasil com o auxílio e utilização de Redes Neurais Artificais (RNAs). Assim, o objetivo geral do artigo foi construir, treinar e validar um modelo de RNAs, com objetivo específico de verificar as correlações entre as variáveis selecionadas com significativa segurança. O software ...