CODEWITHSUNDEEP
pythonmachine-learningscikit-learn
zzzbbx
zzzbbx

Reputation: 10131

Using pipeline with sklearn

I'm trying to define a quantizer to use with Pipeline/GridSearchCV in sklearn. When defining as below

class Quantizer(base.BaseEstimator, base.TransformerMixin):
    def __init__(self):

    def transform(X, y=None):
      some code

I'm getting something like

method fit is missing

Am I missing something in the definition of the class?

Upvotes: 3

Views: 1507

Answers (1)

elyase
elyase

Reputation: 40963

If you are only transforming data in an intermediate state of your pipeline you don't need to implement a fit method, so you just return self:

class Quantizer(base.BaseEstimator, base.TransformerMixin):
    def __init__(self):

    def transform(self, X, y=None):
      # some code 

    def fit(self, X, y=None, **fit_params):
      return self

Take a look here for more details.

Upvotes: 8

Related Questions