Time Series Analysis

Abstract

Recall that in the previous chapter objective or quantitative forecasting methods were defined as forecasting methods that rely on a formalized underlying model to make predictions. They are further divided into time series and causal methods. Time series analysis is a forecasting method based on the fundamental assumption that future estimates are based on prior, historical values of the same variable. This implies that the historical pattern exhibited by the variable to be forecasted will extend into the future. In addition, it is implicitly assumed that historical data are available. The only independent variable in a forecasting model based on time series analysis is the time period. Time series forecasting methods are mostly used to forecast variables for the short to intermediate term. As such, time series methods are some of the forecasting techniques most often used by logisticians and are developed in further detail in this chapter.