Preparing multiple time series for a global model

Now, it is time to move on to the type of time series problems that involve multiple time series. In this recipe, we will learn the fundamentals of global forecasting models and how they work. We’ll also explore how to prepare a dataset that contains multiple time series for forecasting. Again, we leverage the capabilities of the TimeSeriesDataSet and DataModule classes to help us do this.

Getting ready

So far, we’ve been working with time series problems involving a single dataset. Now, we’ll learn about global forecasting models, including the following:

• Transitioning from local to global models: Initially, our work with time series forecasting focused on single datasets, where models predict future values based on historical data of one series. These so-called local models are tailored to specific time series, whereas global models involve handling multiple related time series and capturing relevant...