Variadic function definitions¶
Experimental API for updated model function signatures
As part of a move to make Machine Learning functionality easier to use within the KX libraries, some sections of the ML-Toolkit, and newly added functions≤ have been rewritten using an experimental variadic function definition API. The change to function signatures is to reduce the barrier to entry for users who are new to machine learning and wish to make use of the provided ML functionality using specified default parameters.
A variadic function takes a variable number of parameters and varies its behavior accordingly.
A variadic function is overloaded on rank.
In the context of the ML-Analytics library a variadic function is expected to
- Take a set number of required parameters based on the model, i.e. features and target for supervised models, or just features for unsupervised models
- Optionally have additional configuration that can be set by a user to tune the performance and expected behaviour of a model
This is an experimental API intended as an entry point for developers new to machine learning and q.
The success or otherwise of this form of API will determine its future use. More generally, changes to this API should be expected, including support for
- positional arguments
- keyword arguments
- dictionary arguments