Package mekano :: Package ml :: Module multiclassifier :: Class MultiClassifier

Class MultiClassifier

mc = MultiClassifier()

Manages a set of classifiers for a multi-class/multi-label classification problem. There are two ways to create a MultiClassifier. First, manually create the classifier for each label, then add it to MultiClasifier:

>>> mc.add(label, classifier)

Each classifier should provide a score function.

scores = mc.score(av) 'scores' is a map of label:score

Second, provide a BaseClassifier class and a multi-labeled dataset and optimal params to be passed to BaseClassifier, in which case MultiClassifier with binarize the dataset and do the training:

>>> mc = mekano.MultiClassifier.create(mekano.LogisticRegressionClassifier, trainset, LAMBDA=0.1, c=1.0)

Instance Methods

[hide private]

__init__(self)

source code

add(self, label, classifier)

source code

__getitem__(self, key)

source code

score(self, av)
Score a vector av

source code

__repr__(self)

source code

Static Methods

[hide private]

create(BaseClassifier, ds, **params)
Create a MultiClassifier from a base classifier class and multi-labeled dataset.

source code

Method Details

[hide private]

score(self, av)

source code

Score a vector av

Returns a map of label:score

create(BaseClassifier, ds, **params)
Static Method

source code

Create a MultiClassifier from a base classifier class and multi-labeled dataset.

'params' are optional parameters to pass to the BaseClassifier constructor.

Class MultiClassifier

score(self, av)

create(BaseClassifier, ds, **params) Static Method

create(BaseClassifier, ds, **params)
Static Method