5afc43254eb6d03b3b0ff75bc0ab9a2be621c147,tslearn/nonmyopic.py,NonMyopicEarlyClassification,fit,#NonMyopicEarlyClassification#Any#Any#,96

Before Change

        
        classes_y = np.unique(Y)
        self.cluster_ = TimeSeriesKMeans(n_clusters=self.n_clusters)
        self.classifier_ = []
        self.clusters_ = {}
        self.pyhatyck_ = {}
        self.indice_ck_ = []
        self.__n_classes_ = len(classes_y)
        self.__len_X_ = X.shape[1]
        c_k = self.cluster_.fit_predict((to_time_series_dataset(X)))
        mid_X = int(X.shape[0] / 2)
        X1 = X[:mid_X]
        Y1 = Y[:mid_X]
        X2 = X[mid_X:]
        Y2 = Y[mid_X:]
        c_k2 = c_k[mid_X:]
        vector_of_ones = np.ones((len(X[:]),))
        self.pyck_ = coo_matrix((vector_of_ones, (Y, c_k)), shape=(self.__n_classes_, self.n_clusters)).toarray()
        for k in range(0, self.n_clusters):
            self.clusters_["ck_cm{0}".format(k)] = []
            self.pyhatyck_["pyhatycks{0}".format(k)] = []
            self.indice_ck_.append(np.where(c_k == k))
            current_sum = self.pyck_[:, k].sum()
            for current_classe in range(0, self.__n_classes_):
                self.pyck_[current_classe, k] = self.pyck_[current_classe, k] / current_sum

        for t in range(self.minimum_time_stamp, X.shape[1] + 1):
            self.classifier_.append(
                self.classifier(
                    solver=self.solver,
                    hidden_layer_sizes=self.hidden_layer_sizes,
                    random_state=self.random_state,
                    max_iter=self.maximum_iteration,
                )
            )
            self.classifier_[-1].fit(X1[:, :t], Y1)
            for k in range(0, self.n_clusters):
                index_cluster = np.where(c_k2 == k)
                if len(index_cluster[0]) != 0:

After Change

        
        classes_y = np.unique(Y)
        self.cluster_ = TimeSeriesKMeans(n_clusters=self.n_clusters)
        self.classifiers_ = {t: clone(self.base_classifier)
                             for t in range(self.minimum_time_stamp,
                                            X.shape[1] + 1)}
        self.clusters_ = {}
        self.pyhatyck_ = {}

Instances