Análise de regressão e simulação no gerenciamento de indicadores de desempenho industrial

Abstract

The application of statistical procedures. This research aims to associate two of the most influents stochastic models, regression analysis e systems simulation, in order to facilitate the practical understanding of data analysis. The algorithms are designed on Excel, by using several statistical techniques such as regression theory, ANOVA, Monte Carlo method, and Cholesky Factorization to create models to analyze either simple or multiple systems with up to five independent variables. In this study, the models are tested through analysis performed on industrial performance measures. Among the results, indexes were created to improve the measurement system management, by identifying instability, correlation, and anomalies. Even though the models are simple and accessible, they provided very satisfactory results with numerous possibilities of industrial and academic applications, besides the potential for new analysis techniques.