Time Series Forecasting¶
Perform time series forecasting in Amazon SageMaker
SageMaker Algorithm |
CPU Training |
GPU Training |
Multi-GPU Training |
Incremental Training |
|---|---|---|---|---|
RNN SageMaker Algorithm¶
The RNN SageMaker Algorithm performs time series forecasting with Recurrent Neural Networks (RNNs). The model consists of a stack of RNN layers with either LSTM or GRU cells. Each RNN layer is followed by an activation layer and a dropout layer. The model is trained by minimizing the negative Gaussian log-likelihood and outputs the predicted mean and standard deviation at each future time step. The algorithm can be used for both univariate and multivariate time series and supports the inclusion of external features. For additional information, see the algorithm's AWS Marketplace listing page and GitHub repository.
LNN SageMaker Algorithm¶
The LNN SageMaker Algorithm performs time series forecasting with Liquid Neural Networks (LNNs). The algorithm uses the closed-form continuous-depth (CfC) implementation of LNNs. CfCs use an approximate closed-form solution of the Liquid Time Constant (LTC) Ordinary Differential Equation (ODE) and, as a result, provide faster training and inference performance than other LNNs. The model is trained by minimizing the negative Gaussian log-likelihood and outputs the predicted mean and standard deviation at each future time step. The algorithm can be used for both univariate and multivariate time series and supports the inclusion of external features. For additional information, see the algorithm's AWS Marketplace listing page and GitHub repository.