Skip to content


Base class that is inherited by the majority of classes in River.

This base class allows us to handle the following tasks in a uniform manner:

  • Getting and setting parameters

  • Displaying information

  • Mutating/cloning



Return a fresh estimator with the same parameters.

The clone has the same parameters but has not been updated with any data. This works by looking at the parameters from the class signature. Each parameter is either - recursively cloned if its a class. - deep-copied via copy.deepcopy if not. If the calling object is stochastic (i.e. it accepts a seed parameter) and has not been seeded, then the clone will not be idempotent. Indeed, this method's purpose if simply to return a new instance with the same input parameters.


  • new_params'dict | None' — defaults to None
  • include_attributes — defaults to False


Modify attributes.

This changes parameters inplace. Although you can change attributes yourself, this is the recommended way to proceed. By default, all attributes are immutable, meaning they shouldn't be mutated. Calling mutate on an immutable attribute raises a ValueError. Mutable attributes are specified via the _mutable_attributes property, and are thus specified on a per-estimator basis.


  • new_attrs'dict'