Evaluation of insolvency in cooperative credit unions: An application

The present study aimed to analyze the state of insolvency of mutual cooperative credit unions in the Parana State, Brazil, building a mathematical model based on Artificial Neural Networks (ANN), whose main feature is the ability to process information, using factors of brain functioning such as the ability to learn, decide and adapt to changes as inspiration. The information needed to build the model was obtained from a sample of 62 mutual credit unions (31 solvent ones and 31 insolvent) by calculating the financial indicators of the PEARLS system. Algorithms of RBFNetwork, MultilayerPerceptron and MultilayerPerceptronCS neural networks, from Weka Software Package, a tool utilized in Data Mining Tool, were used to model 27, 10 and 11 indicators. The ANN with MultilayerPerceptron and MultilayerPerceptronCS algorithms modeling 27 indicators showed a better performance when compared with other built networks. Keywords: Artificial Neural Networks, Credit Union, Insolvency; Regional Development.