Credit Risk Analysis Model in Microfinance Institutions in Peru Through the use of Bayesian Networks

In this paper, we propose a risk analysis model to obtain the probability of default of microfinance clients in Peru. Our model uses trends of predictive analysis through variants of neural network algorithms; and data processing methodologies such as the Knowledge Discovery in Databases (KDD). The analysis method is used through Bayesian networks which allows the customer data evaluation and is related to our model data. This model is composed of 5 phases: 1. The input elements for the analysis; 2. The process of evaluation and analysis; 3. The regulatory standards; 4. The technological architecture; 5. The output elements. This model allows knowing the probability of compliance of a client with 84% prediction accuracy. The model validation was carried out in a microfinance institution in Lima, Peru, using cross-validation, evaluating the sensitivity and specificity of the results.