def fit(self,
data,
class_name,
categorial_name,
multidimentionality=False):
# X,y series
self.categorial_name = categorial_name
conjugate_matrix = ConjugateMatrix()
conjugate_matrix.make(data, class_name, categorial_name)
probability_matrix = ProbabilityMatrix()
probability_matrix.make_form_cobjugate_matrix(conjugate_matrix)
conjugate_matrix.get_D1()
# conjugate_matrix = None
D1 = probability_matrix.get_D1()
D2 = probability_matrix.get_D2()
D1_sqrt_ = fractional_matrix_power(D1, -0.5)
D2_sqrt_ = fractional_matrix_power(D2, -0.5)
PH = np.dot(np.dot(D1_sqrt_, probability_matrix.matrix), D2_sqrt_)
T_1 = np.dot(PH.T, PH)
# T_2 = np.dot(PH,PH.T)
tags = Tags()
tags.make(T_1, D2_sqrt_, multidimentionality)
self.C = tags.tags
return self
def fit(self, X, multidimentionality=False):
X = deepcopy(X)
# X - only caregorial
self.columns = X.columns.values
p = X.columns.values.shape[0]
binary_matrix = BinaryMatrixData()
binary_matrix = binary_matrix.make(X)
D = binary_matrix.make_D()
D_sqrt_ = fractional_matrix_power(D, -0.5)
D = None
F = np.dot(((1 / p ** 0.5) * binary_matrix.matrix), D_sqrt_)
binary_matrix = None
T_1 = np.dot(F.T, F)
F = None
tags = Tags()
tags.make(T_1,D_sqrt_,multidimentionality)
T_1 = None
self.C = tags.tags
return self
