Building time series forecasting models and understanding their model outcomes
Similar to projects we looked at in Chapter 4, Preparing Data for DataRobot, through to Chapter 8, Model Scoring and Deployment, once we have finished with the initial configurations, we scroll up and click on the Start button. By doing this, DataRobot automatically builds time series models for this project. Before we evaluate the models, it would be useful to understand the nature of the features the platform extracts. DataRobot extracts features from the data that differ considerably from those of other prediction models, as is evident in the following screenshot:
Figure 9.4 – Feature lists
The lists shown under the Feature Lists tab are constructed as part of exploratory data analysis (EDA) and itemize differing lists of features that DataRobot employs in creating models. Many of the feature lists involve derived features, which are created automatically based on properties...
