Skip to content

Latest commit

 

History

History
97 lines (66 loc) · 3.26 KB

Orange.classification.rst

File metadata and controls

97 lines (66 loc) · 3.26 KB
.. automodule:: Orange.classification

Classification (classification)

Induction of models in Orange is implemented through a two-class schema. A learning algorithm is represented by an instance of a class derived from :obj:`Orange.classification.Learner`. The learner stores all parameters of the learning algorithm. Induced models are represented by instances of classes derived from :obj:`Orange.classification.Classifier`.

Therefore, to induce models from data, one first needs to construct the instance representing a learning algorithm (e.g. :obj:`~Orange.classification.tree.TreeLearner`) and set its parameters. Calling the learner with some training data returns a classifier (e.g. :obj:`~Orange.classification.tree.TreeClassifier`). The learner does not "learn" to classify but constructs classifiers.

.. literalinclude:: code/bayes-run.py
   :lines: 7-

To simplify the procedure, the learner's constructor can also be given training data, in which case it fits and returns a model (an instance of :obj:`~Orange.classification.Classifier`) instead of a learner:

classifier = Orange.classification.bayes.NaiveLearner(titanic)

Orange contains a number of learning algorithms described in detail on separate pages.

.. toctree::
   :maxdepth: 1

   Orange.classification.bayes
   Orange.classification.knn
   Orange.classification.rules
   Orange.classification.svm
   Orange.classification.tree
   Orange.classification.logreg
   Orange.classification.majority
   Orange.classification.lookup
   Orange.classification.classfromvar
   Orange.classification.constant
   Orange.classification.neural

All learning algorithms and prediction models are derived from the following two clases.

Abstract base class for learning algorithms.

.. method:: __call__(data[, weightID])

    An abstract method that fits a model and returns it as an
    instance of :class:`Classifier`. The first argument gives the
    data (as :obj:`Orange.data.Table` and the optional second
    argument gives the id of the meta attribute with instance
    weights.

Abstract base class for prediction models (both classifiers and regressors).

.. method:: __call__(instance, return_type=GetValue)

    Classify a new instance using this model. Results depends upon
    the second parameter that must be one of the following.

    :obj:`Orange.classification.Classifier.GetValue`

        Return value of the target class when performing prediction.

    :obj:`Orange.classification.Classifier.GetProbabilities`

        Return probability of each target class when performing prediction.

    :obj:`Orange.classification.Classifier.GetBoth`

        Return a tuple of target class value and probabilities for each class.


    :param instance: data instance to be classified.
    :type instance: :class:`~Orange.data.Instance`

    :param return_type: what needs to be predicted
    :type return_type: :obj:`GetBoth`,
                       :obj:`GetValue`,
                       :obj:`GetProbabilities`

    :rtype: :class:`~Orange.data.Value`,
          :class:`~Orange.statistics.distribution.Distribution` or a
          tuple with both