Especificação e implementação de um sistema evolutivo de apoio à  análise de demonstrações contábeis SEADC para fins de tomada de decisão

• Ano de defesa: 2012
• Autor(a) principal: Neves, Cleonábula Maria Maranhão
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Alagoas; BR; Modelagem Computacional de Conhecimento; Programa de Pós-Graduação em Modelagem Computacional de Conhecimento; UFAL
• Palavras-chave em Português: Evolutionary computation; Genetic algorithm; Accounting analysis; Computational modeling; Economic and financial indicators; Computação evolutiva; Algoritmo genético; Modelagem computacional; Análise contábil; Indicadores econômicos e financeiros; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: http://repositorio.ufal.br/handle/riufal/851

This paper describes the specification and implementation of a computational system entitled Evolutionary Support System for the Analysis of Accounting Statements (ESAAS), which aims to optimize decision making, and uses variations of the genetic algorithm from Holland. In its specification, there is a population formed by chromosomes that are vectors over a ternary alphabet with size corresponding to the number of account statements contained in the type of analysis chosen by the user. The ESAAS diagnoses the financial and economic situation of commercial, industrial companies or service providers, whether they are classified as micro, small, average, average-large or large companies. The diagnosis done by ESAAS was based on the analyzes of liquidity, debt profitability, added value, average terms, vertical and horizontal analyzes; and in the diagnosis of the analytical combination of liquidity, profitability and debt with the vertical analysis, showing as a justification, the influence of each of the accounts or group of financial accounts involved. The system receives as input a set of indexes, total values of the accounts or groups of accounts of financial accounting standards of the company that will be considered in a determinted period, consisting of consecutive years, being considered the minimum of three years. To validate the ESAAS, tests were performed with data obtained from the financial statements from companies in the years 2008, 2009 and 2010. The system was implemented in the eclipse compiler, using the JAVA programming language. The results achieved in the calculation of the quotients obtained in all the seven types of analyzes performed achieved 100% of correctness. Therefore concluding that the ESAAS infers knowledge to decision-making, through diagnostic reports generated.