Training and deploying a DeepAR model
The goal of forecasting models is to predict future data points based on previous records. There are different forecasting algorithms available, including ARIMA and ETS. One algorithm making use of recurrent neural networks (RNNs) to forecast time series data is DeepAR. In this recipe, we will train and deploy a DeepAR model using the SageMaker Python SDK. To help us get started with using the built-in DeepAR forecasting algorithm, we will only work with a single time series dataset when training the model.
Getting ready
Here are the prerequisites of this recipe:
- This recipe continues from Performing the train-test split on a time series dataset.
- A SageMaker Studio Notebook running the Python 3 (Data Science) kernel.
How to do it…
The first few steps in this recipe focus on preparing the prerequisites for training the DeepAR model:
- Create a new notebook using the
Python 3 (Data Science)kernel inside...
