Note: The translation for this entry is currently under quality review. Some content is temporarily displayed in English only.
hyperparameter
This term is specific to the domain of machine learning and computational statistics. It describes the external settings that govern the training process, distinguishing them from parameters, which are the internal weights the model learns automatically from the data.
In a technical context, hyperparameters are often subject to "tuning" or "optimization," where a developer experiments with different values to find the most efficient configuration for a specific dataset. This process is distinct from the actual training phase of the model.
Meanings
A configuration setting in a machine learning model that is specified before the training process begins and remains constant during training, controlling the overall behavior of the learning algorithm.
The data scientist adjusted the learning rate hyperparameter to prevent the model from overshooting the minimum loss.