Univariate Time Series Analysis: Box Jenkins Methodology. Examples and Exercises with SAS

The Box–Jenkins methodology is the classical framework for time series modeling, especially for ARIMA models. The Box–Jenkins approach provides a systematic process for identifying, estimating, and checking ARIMA (AutoRegressive Integrated Moving Average) models. This book develops univariate time series analysis using the Box-Jenkins methodology. It begins by covering concepts such as the components of a time series, the decomposition of a series into its components, stationarity, seasonality, and autocorrelation and partial autocorrelation functions. Non-seasonal and seasonal deterministic models are then studied, considering the Holt, Brown, Winters, and other exponential models. It then delves into non-seasonal and seasonal ARIMA models according to the Box-Jenkins methodology, considering the AR, MA, ARMA, and ARIMA models. Automatic prediction methods are also considered. All concepts are illustrated with fully implemented examples and exercises using SAS software, which is very suitable for time series analysis.