Book Description

This book develops the Box and Jenkins methodology for the prediction of time series through the ARIMA models. The book begins by introducing the concepts needed to make univariate time series predictions. Next, the identification, estimation and prediction of the ARIMA models is deepened, both in the non-seasonal field and in the seasonal field. An important part of the content is the automatic prediction methods, including the use of neural networks and the space of the states to obtain improved predictions of time series. The intervention models that collect the effects of atypicalities in obtaining predictions are discussed below. Finally, the transfer function models or ARIMAX models that use external continuous regressors to guide the predictions of a time series are considered. A great variety of examples and exercises solved with R. are presented.