In this chapter, we introduced the forecasting applications of the linear regression model. Although the linear regression model was not designed to handle time series data, with simple feature engineering we can transform a forecasting problem into a linear regression problem. The main advantage of the linear regression model with respect to other traditional time series models is the ability of the model to incorporate external variables and factors. Nevertheless, this model can handle time series with multiseasonality patterns, as we saw with the UK demand for electricity forecast. Last but not least, the forecasting approaches we demonstrated in this chapter will be the base for advanced modeling with machine learning models that we will discuss in Chapter 12, Forecasting with Machine Learning Models.
In the next chapter, we will introduce the exponential smoothing...