summit.multiview_platform.monoview_classifiers.scm

scm

classifier_class_name = 'SCM'
class SCM(random_state=None, model_type='conjunction', max_rules=10, p=0.1, **kwargs)

SCM Classifier :param random_state (default: :type random_state (default: None) :param model_type: :type model_type: string (default: “conjunction”) :param max_rules: :type max_rules: int number maximum of rules (default : 10) :param p: :type p: float value(default : 0.1 ) :param kwarg: :type kwarg: others arguments

param_names
distribs
classed_params
weird_strings
param_names = ['model_type', 'max_rules', 'p', 'random_state']
distribs
classed_params = []
weird_strings
fit(X, y, tiebreaker=None, iteration_callback=None, **fit_params)

Fit a SCM model.

Parameters:

X: array-like, shape=[n_examples, n_features]

The feature of the input examples.

yarray-like, shape = [n_samples]

The labels of the input examples.

tiebreaker: function(model_type, feature_idx, thresholds, rule_type)

A function that takes in the model type and information about the equivalent rules and outputs the index of the rule to use. The lists respectively contain the feature indices, thresholds and type corresponding of the equivalent rules. If None, the rule that most decreases the training error is selected. Note: the model type is provided because the rules that are added to disjunction models correspond to the inverse of the rules that are handled during training. Handle this case with care.

iteration_callback: function(model)

A function that is called each time a rule is added to the model.

Returns:

self: object

Returns self.

get_interpretation(directory, base_file_name, y_test, feature_ids, multi_class=False)

Base method that returns an empty string if there is not interpretation method in the classifier’s module

paramsToSet(nIter, random_state)