def calcualteTDW(self, row, groupColumns):
data = [row[col].values.astype(np.float) for col in groupColumns]
#data = [x[~np.isnan(x)] for x in data]
distMatrix = cdist(data[0].reshape(len(groupColumns[0]), 1),
data[1].reshape(len(groupColumns[1]), 1))
return SoftDTW(distMatrix).compute()
def _softdtw_func(Z, X, weights, barycenter, gamma):
# Compute objective value and grad at Z.
Z = Z.reshape(barycenter.shape)
G = numpy.zeros_like(Z)
obj = 0
for i in range(len(X)):
D = SquaredEuclidean(Z, X[i])
sdtw = SoftDTW(D, gamma=gamma)
value = sdtw.compute()
E = sdtw.grad()
G_tmp = D.jacobian_product(E)
G += weights[i] * G_tmp
obj += weights[i] * value
return obj, G.ravel()
def _func(self, Z):
# Compute objective value and grad at Z.
Z = Z.reshape(self.barycenter_.shape)
G = numpy.zeros_like(Z)
obj = 0
for i in range(len(self._X_fit)):
D = SquaredEuclidean(Z, self._X_fit[i])
sdtw = SoftDTW(D, gamma=self.gamma)
value = sdtw.compute()
E = sdtw.grad()
G_tmp = D.jacobian_product(E)
G += self.weights[i] * G_tmp
obj += self.weights[i] * value
return obj, G.ravel()
