.. _`Pipeline`:

.. _`org.sysess.sympathy.machinelearning.pipeline`:

Pipeline
````````

.. image:: pipeline.svg
   :width: 48


Applies one model on the output of another




Documentation
:::::::::::::

Creates a sequence of data transformers with an optional final predictor.

Intermediate steps of the pipeline must be transformers, that is, they must implement fit and
transform methods. The final estimator only needs to implement fit. The transformers in the
pipeline can be cached using memory argument. The pipline serves multiple purposes:

    - It assembles several steps that can be cross-validated together while setting different
      parameters.
    - Only one fit/predict call for the whole sequence.
    - Avoid statistics leaking from test to train in cross-validation.


Definition
::::::::::


Input ports
...........

    **models**
        | Type: model
        | Description: models
        | Optional number of ports: 2--inf (default: 2)
        

Output ports
............

    **out-model**
        | Type: model
        | Description: Output model



Configuration
.............

    **Flatten** (flatten)
        Flattens multiple pipeline objects into a single pipeline containing all models
    **Model names** (names)
        Comma separated list of model names, eg. Rescale, SVC. If fewer names are given than models then default names will be used.







Examples
........





The node can be found in:

* :download:`Custom_ParameterSearch.syx </Examples/Sympathy for Data/MachineLearning/Custom_ParameterSearch.syx>`

* :download:`Hyper_parameters.syx </Examples/Sympathy for Data/MachineLearning/Hyper_parameters.syx>`

* :download:`Eigenfaces.syx </Examples/Sympathy for Data/MachineLearning/Eigenfaces.syx>`

* :download:`Polynomial_Features.syx </Examples/Sympathy for Data/MachineLearning/Polynomial_Features.syx>`




Implementation
..............

.. automodule:: node_pipeline
    :noindex:

.. class:: Pipeline
    :noindex: