def predict(self,
xdata=None,
spline_parameters=None,
df=None,
spline_type=None,
*args,
**kwargs):
if xdata is None:
xdata = self.xdata
if spline_parameters is None:
if self.spline_parameters is None:
warn(
"Spline parameters are not specified, you should either fit a model or specify them. "
"Output is set to 0")
return np.zeros((xdata.shape[0], 1))
else:
spline_parameters = self.spline_parameters
if df is None:
if self.df is None:
warn(
"Splines df is not specified, you should either fit a model or specify them. "
"Output is set to 0")
return np.zeros((xdata.shape[0], 1))
else:
df = self.df
if spline_type is None:
if self.spline_type is None:
warn(
"Spline spline_type are not specified, you should either fit a model or specify them. "
"Output is set to 0")
return np.zeros((xdata.shape[0], 1))
else:
spline_type = self.spline_type
fit_wrapper = LR(fit_intercept=False)
fit_wrapper.coef = spline_parameters
if spline_type == RegressionSplinesSmoother._spline_type_list.index(
'bs'):
y_pred = np.dot(
dmatrix("bs(predictor, df=" + str(self.df) +
", include_intercept=False)", {"predictor": xdata},
return_type='dataframe'), spline_parameters)
elif spline_type == RegressionSplinesSmoother._spline_type_list.index(
'cr'):
y_pred = np.dot(
dmatrix("cr(predictor, df=" + str(self.df) + ")",
{"predictor": xdata},
return_type='dataframe'), spline_parameters)
else:
raise ValueError('Please, specify a valid splines_type')
return y_pred
