Making predictions with time series models
DataRobot provides us with tools to make predictions pain-free. There are two approaches to making predictions for time series. For small datasets under 1 gigabyte (GB), predictions could be made using the Make Predictions tab on the Leaderboard feature. This involves setting up and uploading a prediction dataset, then scoring it within the Drag and drop a new dataset user interface (UI) functionality. For significantly larger datasets, models need to be deployed and predictions are made using an application programming interface (API). In this chapter, we will cover the first approach to making predictions. With DataRobot, general model deployments and working with APIs are extensively discussed in Chapter 12, DataRobot Python API.
The leaderboard's drag-and-drop approach to scoring models for time series models somewhat differs from those of traditional models, as seen in Chapter 8, Model Scoring and Deployment. When the Make Predictions...
