#window_length=200 #overlap_points=185 #kt=2 #kf=2 #perf=69.04% tol = np.power(10, -6, dtype=float) maximum_iterations = 10 C_values = np.power(10, np.linspace(-2, 9, 6), dtype=float) #C_values=np.linspace(0.001,10000,20) G_values = np.power(10, np.linspace(-4, -1, 4), dtype=float) Hyper_values = np.array([np.power(10, -8, dtype=float)]) #10^(-2) C_svm, G_svm, hyperparam, matrixscore = Methods.Cross_valHALs( Tensor_train, y_train, C_values, G_values, Hyper_values, nbclasses, kf, kt, maximum_iterations, tol) #We perform the decomposition of the training tensor #The decomposition yields: #The error related to each updating error_list; #The temporal and spectral dictionary components A_f and A_t; #The number of iterations and the activation coefficients G; #We dimension purpose, we can reduce the size of the tensor to be decomposed.This can be done in the following way: #Tensor_train=dtensor(Preprocessing.rescaling(Tensor_train,If,It)) where If and It correspond to the desired sizes. G, A_f, A_t, error_list = Methods.PenalizedTuckerHals( Tensor_train, y_train, nbclasses, kf, kt, hyperparam, maximum_iterations, tol) # #We define the training features. They are obtained by vectorizing the matrices G[k,:,:] #Training_features=Methods.Training_features_extraction(G)