Mínimos cuadrados inmersos en la predicción de presupuesto en cobranzas

Abstract

For the Collections area, it is important to have control over the prediction of data, especially in relevant data such as the Non-performing Portfolio, when considering methods such as Lagrange polynomials, the result may fail due to the Runge phenomenon. This phenomenon refers to excessive oscillation and fluctuation of the interpolated polynomials as the degree increases. This erratic behavior can lead to the prediction of inaccurate results. On the other hand, if the prediction by means of least squares is considered, a more successful solution for the prediction of data is demonstrated. In this method, the data curve is fitted to a polynomial function through a process of minimizing square errors.