Forecasting and Time Series Models
Chapter Preview. This chapter introduces two popular smoothing techniques, moving (running) averages and exponential smoothing, for forecasting. These techniques are simple to explain and easily interpretable. They can be also expressed as regression models, where the technique of weighted least squares is used to compute parameter estimates. Seasonality is then presented, followed by a discussion of two more advanced time series topics, unit root testing, and volatility (ARCH/GARCH) models.
9.1 Smoothing with Moving Averages
Smoothing a time series with a moving, or running, average, is a time-tested procedure. ...