Esempio n. 1
0
class CDEAC(object):

    def __init__(self, base_clf_class):
        self.base_clf_class = base_clf_class


    def fit(self, X_train, y_train):
        self.weight = GenericWeights(self.base_clf_class)

        ac = AC2(self.base_clf_class)
        ac.fit([X_train, y_train])
        self.ac = ac

        self.X_train = X_train
        self.y_train = y_train


    def predict(self, X_test):
        class_dist = self.ac.predict_dict(X_test)
        self.weight.fit(self.X_train, self.y_train, class_dist)
        y_pred = self.weight.predict(X_test)
        return y_pred
Esempio n. 2
0
class CDEITR(object):
    ''' Combination of EM based CDE and Cost sensitive learning based on class
    reweighting. '''

    def __init__(self, base_clf_class):
        self.base_clf_class = base_clf_class


    def fit(self, X_train, y_train):
        self.weight = GenericWeights(self.base_clf_class)

        it = Itr2(self.base_clf_class)
        it.fit([X_train, y_train])
        self.it = it

        self.X_train = X_train
        self.y_train = y_train


    def predict(self, X_test):
        class_dist = self.it.predict_dict(X_test)
        self.weight.fit(self.X_train, self.y_train, class_dist)
        y_pred = self.weight.predict(X_test)
        return y_pred