def periodic_fit(t, y, dy, frequency, t_fit,
center_data=True, fit_bias=True, nterms=1):
"""Compute the Lomb-Scargle model fit at a given frequency
Parameters
----------
t, y, dy : float or array_like
The times, observations, and uncertainties to fit
frequency : float
The frequency at which to compute the model
t_fit : float or array_like
The times at which the fit should be computed
center_data : bool (default=True)
If True, center the input data before applying the fit
fit_bias : bool (default=True)
If True, include the bias as part of the model
nterms : int (default=1)
The number of Fourier terms to include in the fit
Returns
-------
y_fit : ndarray
The model fit evaluated at each value of t_fit
"""
if dy is None:
dy = 1
t, y, dy = np.broadcast_arrays(t, y, dy)
t_fit = np.asarray(t_fit)
assert t.ndim == 1
assert t_fit.ndim == 1
assert np.isscalar(frequency)
if center_data:
w = dy ** -2.0
y_mean = np.dot(y, w) / w.sum()
y = (y - y_mean)
else:
y_mean = 0
X = design_matrix(t, frequency, dy=dy, bias=fit_bias, nterms=nterms)
theta_MLE = np.linalg.solve(np.dot(X.T, X),
np.dot(X.T, y / dy))
X_fit = design_matrix(t_fit, frequency, bias=fit_bias, nterms=nterms)
return y_mean + np.dot(X_fit, theta_MLE)
评论列表
文章目录
