Leveraging HPO
Training an ML model is a process that consists of finding parameters that will help the model to better deal with real data. When you train your own model without using a managed service such as Amazon Forecast, you can encounter three types of parameters, as follows:
- Model selection parameters: These are parameters that you have to fix to select a model that best matches your dataset. In this category, you will find the
a
,b
,g
, andj
parameters from the ETS algorithm, for instance. Amazon Forecast implements these algorithms to ensure that automatic exploration is the default behavior for ETS and ARIMA so that you don't have to deal with finding the best values for these by yourself. For other algorithms (such as NPTS), good default parameters are provided, but you have the flexibility to adjust them based on the inner knowledge of your datasets. - Coefficients: These are values that are fitted to your data during the very training of your model. These...