def __init__(self, x, y, yerr, subdivisions):
"""
Initialize global variables of Gaussian Process Regression Interpolator
Args:
x (array): Independent variable
y (array): Dependent variable
yerr (array): Uncertainty on y
subdivisions: The number of subdivisions between data points
"""
# Define kernels
kernel_expsq = 38**2 * kernels.ExpSquaredKernel(metric=10**2)
kernel_periodic = 150**2 * kernels.ExpSquaredKernel(
2**2) * kernels.ExpSine2Kernel(gamma=0.05, log_period=np.log(11))
kernel_poly = 5**2 * kernels.RationalQuadraticKernel(
log_alpha=np.log(.78), metric=1.2**2)
kernel_extra = 5**2 * kernels.ExpSquaredKernel(1.6**2)
kernel = kernel_expsq + kernel_periodic + kernel_poly + kernel_extra
# Create GP object
self.gp = george.GP(kernel, mean=np.mean(y), fit_mean=False)
self.gp.compute(x, yerr)
# Set global variables
self.ndim = len(self.gp)
self.x = x
self.y = y
self.yerr = yerr
self.subdivisions = subdivisions
self.priors = [prior.Prior(0, 1) for i in range(self.ndim)]
self.x_predict = np.linspace(min(self.x), max(self.x),
subdivisions * (len(self.x) - 1) + 1)
def create_data(gamma=10, period=1, amp=1):
"""Create a randomized periodic-sinusoidal function.
Some example inputs are:
gamma = 10
period = 1
amp = 1
Returns all x-values with corresponding y-values and gp
function generated with the given gamma, period, and amplitude.
"""
#generate 10 random numbers for x
pre_x = 10 * np.sort(np.random.rand(10))
#determine the y error and make it an array as long as pre_x
#have the error scale with the amplitude
pre_yerr = 0.2 * amp * np.ones_like(pre_x)
x_possible = np.linspace(0, 10, 1000)
#create a sinusoidal plot based on inputs
#establish the kernel type we are using
kernel = amp * kernels.ExpSine2Kernel(gamma, period)
gp = george.GP(kernel)
#generate a ExpSine2 function using our x-values (0-10) and our y error
gp.compute(pre_x, pre_yerr)
#note: pre_x is never used again after this
#sample the y-values from the function we made
pre_y = gp.sample(x_possible) #a subset of possible x-values
#return our simulated data with the original function
return (x_possible, pre_y, gp)
def kern_QP(p):
nper = (len(p) - 1) / 4
for i in range(nper):
log_a, log_gamma, period, log_tau = p[i * 4:i * 4 + 4]
a_sq = 10.0**(2 * log_a)
gamma = 10.0**log_gamma
tau_sq = 10.0**(2 * log_tau)
if i == 0:
kern = a_sq * kernels.ExpSine2Kernel(gamma, period) * \
kernels.ExpSquaredKernel(tau_sq)
else:
kern += a_sq * kernels.ExpSine2Kernel(gamma, period) * \
kernels.ExpSquaredKernel(tau_sq)
log_sig_ppt = p[-1]
sig = 10.0**(log_sig_ppt - 3)
return kern, sig
def computeModel():
global gp
kernel = np.var(y) * kernels.ExpSquaredKernel(0.5) * kernels.ExpSine2Kernel(log_period = 0.5, gamma=1)
gp = george.GP(kernel)
gp.compute(x, y)
model = gp.predict(y, x, return_var=True)
result = minimize(neg_ln_like, gp.get_parameter_vector(), jac=grad_neg_ln_like)
gp.set_parameter_vector(result.x)
def __init__(self, P=None, *args, **kwargs):
"""Initialises the priors, initial hp values, and kernel."""
# Set up kernel
# -------------
k_spatial = 1.0 * kernels.ExpSquaredKernel(
metric=[1.0, 1.0], ndim=3, axes=[0, 1])
k_temporal = 1.0 * kernels.ExpSquaredKernel(
metric=1.0,
ndim=3, axes=2) \
* kernels.ExpSine2Kernel(
gamma=2, log_period=1,
ndim=3, axes=2)
k_total = k_spatial + k_temporal
if P is None:
P = 0.5
# NOTE: sigt always starts as multiple of the period
super().__init__(kernel=k_total,
parameter_names=('ln_Axy', '2ln_sigx', '2ln_sigy',
'ln_At', '2ln_sigt', 'gamma', 'lnP'),
default_values=(-12.86, -3.47, -4.34, -12.28,
2 * np.log(4 * P), 1.0, np.log(P)),
*args,
**kwargs)
# self.set_hyperpriors(keyword=keyword)
# self.set_bounds(keyword=keyword)
# self.parameter_names = ('ln_Axy', '2ln_sigx', '2ln_sigy',
# 'ln_At', '2ln_sigt', 'gamma', 'lnP')
# self.default_values = (-12.86, -3.47, -4.34, -12.28,
# max(2*np.log(4*P), self.bounds[4][0] + 1e-6),
# 1.0, np.log(P))
# self.kernel = k_total
# self.kernel.set_parameter_vector(np.array(self.default_values))
# # Additional potential tools
# self.get_parameter_vector = self.kernel.get_parameter_vector
# self.set_parameter_vector = self.kernel.set_parameter_vector
if np.log(P) < self.bounds[-1][0] or np.log(P) > self.bounds[-1][1]:
raise PriorInitialisationError(
("Initial period is out of bounds\nperiod: {},\n"
"lnP: {}, \nbounds: {}".format(P, np.log(P),
self.bounds[-1])))
elif not np.isfinite(self.log_prior(self.default_values)):
raise PriorInitialisationError(
("Initial hyperparameter values are out of "
"prior bounds.\n"
"hp_default: {}\n"
"bounds: {}\n"
"P: {}".format(self.default_values, self.bounds, P)))
self.default_X_cols = ['x', 'y', 't']
def test_bounds():
kernel = 10 * kernels.ExpSquaredKernel(1.0, metric_bounds=[(None, 4.0)])
kernel += 0.5 * kernels.RationalQuadraticKernel(log_alpha=0.1, metric=5.0)
gp = GP(kernel, white_noise=LinearWhiteNoise(1.0, 0.1))
# Test bounds length.
assert len(gp.get_parameter_bounds()) == len(gp.get_parameter_vector())
gp.freeze_all_parameters()
gp.thaw_parameter("white_noise:m")
assert len(gp.get_parameter_bounds()) == len(gp.get_parameter_vector())
# Test invalid bounds specification.
with pytest.raises(ValueError):
kernels.ExpSine2Kernel(gamma=0.1, log_period=5.0, bounds=[10.0])
def angus_kernel(theta):
"""
use the kernel that Ruth Angus uses. Be sure to cite her
"""
theta = np.exp(theta)
A = theta[0]
l = theta[1]
G = theta[2]
sigma = theta[4]
P = theta[3]
kernel = (A * kernels.ExpSquaredKernel(l) *
kernels.ExpSine2Kernel(G, P) +
kernels.WhiteKernel(sigma)
)
return kernel
def get_kernel(architecture, kernel_name, domain_name_lst, n_dims):
if architecture == "trans":
mapping = trans_domain_kernel_mapping
else:
mapping = None
kernel = None
initial_ls = np.ones([n_dims])
if kernel_name == "constant":
kernel = kernels.ConstantKernel(1, ndim=n_dims)
elif kernel_name == "polynomial":
kernel = kernels.PolynomialKernel(log_sigma2=1, order=3, ndim=n_dims)
elif kernel_name == "linear":
kernel = kernels.LinearKernel(log_gamma2=1, order=3, ndim=n_dims)
elif kernel_name == "dotproduct":
kernel = kernels.DotProductKernel(ndim=n_dims)
elif kernel_name == "exp":
kernel = kernels.ExpKernel(initial_ls, ndim=n_dims)
elif kernel_name == "expsquared":
kernel = kernels.ExpSquaredKernel(initial_ls, ndim=n_dims)
elif kernel_name == "matern32":
kernel = kernels.Matern32Kernel(initial_ls, ndim=n_dims)
elif kernel_name == "matern52":
kernel = kernels.Matern52Kernel(initial_ls, ndim=n_dims)
elif kernel_name == "rationalquadratic":
kernel = kernels.RationalQuadraticKernel(log_alpha=1,
metric=initial_ls,
ndim=n_dims)
elif kernel_name == "expsine2":
kernel = kernels.ExpSine2Kernel(1, 2, ndim=n_dims)
elif kernel_name == "heuristic":
kernel = mapping[domain_name_lst[0]](ndim=n_dims, axes=0)
for i in range(len(domain_name_lst[1:])):
d = domain_name_lst[1:][i]
kernel += mapping[d](ndim=n_dims, axes=i)
elif kernel_name == "logsquared":
kernel = kernels.LogSquaredKernel(initial_ls, ndim=n_dims)
return kernel
def add_spots(self,QP=[5e-5,0.5,30.,28.]):
"""
This attribute add stellar variability using a Quasi-periodic Kernel
The activity is added using a george Kernel
"""
if not hasattr(self,'flux_spots'):
A = QP[0]
le = QP[1]
lp = QP[2]
P = QP[3]
from george import kernels, GP
k = A * kernels.ExpSine2Kernel(gamma=1./2/lp,log_period=np.log(P)) * \
kernels.ExpSquaredKernel(metric=le)
gp = GP(k)
self.flux_spots = 1 + gp.sample(self.time)
self.flux = self.flux * self.flux_spots
self.spots = True
def __init__(self, P=None, **kwargs):
"""Initialises the priors, initial hp values, and kernel."""
# Set up kernel
k_temporal = 1.0 * kernels.ExpSquaredKernel(metric=1.0, ndim=1) \
* kernels.ExpSine2Kernel(gamma=2, log_period=1, ndim=1)
if P is None:
P = 0.5
# NOTE: sigt always starts as multiple of the period
super().__init__(kernel=k_temporal,
parameter_names=('ln_At', '2ln_sigt', 'g', 'lnP'),
default_values=(-12.28, 2 * np.log(4 * P), 1.0,
np.log(P)),
**kwargs)
if not np.isfinite(self.log_prior(self.default_values)):
raise PriorInitialisationError(
"Initial hyperparameter values are out of prior bounds.")
self.default_X_cols = 't'
def get_gp(self,Kernel="Exp",amplitude=1e-3,metric=10.,gamma=10.,period=10.):
"""
Citlalicue uses the kernels provided by george,
now the options are "Exp", "Matern32", "Matern52", and Quasi-Periodic "QP"
User can modify hyper parameters amplitude, metric, gamma, period.
"""
import george
from george import kernels
if Kernel == "Matern32":
kernel = amplitude * kernels.Matern32Kernel(metric)
elif Kernel == "Matern52":
kernel = amplitude * kernels.Matern52Kernel(metric)
elif Kernel == "Exp":
kernel = amplitude * kernels.ExpKernel(metric)
elif Kernel == "QP":
log_period = np.log(period)
kernel = amplitude * kernels.ExpKernel(metric)*kernels.ExpSine2Kernel(gamma,log_period)
#Save the kernel name as an attribute
self.kernel = kernel
#Compute the kernel with George
self.gp = george.GP(self.kernel,mean=1)
#We compute the covariance matrix using the binned data
self.gp.compute(self.time_bin, self.ferr_bin)
def kernel4(data):
def neg_ln_like(p):
gp.set_parameter_vector(p)
return -gp.log_likelihood(y)
def grad_neg_ln_like(p):
gp.set_parameter_vector(p)
return -gp.grad_log_likelihood(y)
try:
x, y, err = data
ls = get_ls(x, y, err)
k = np.var(y) * kernels.ExpSquaredKernel(ls**2)
k2 = kernels.ExpSquaredKernel(90**2) * kernels.ExpSine2Kernel(
gamma=ls, log_period=ls)
kernel = k + k2
gp = george.GP(kernel,
fit_mean=True,
white_noise=np.max(err)**2,
fit_white_noise=True)
gp.compute(x, err)
results = optimize.minimize(neg_ln_like,
gp.get_parameter_vector(),
jac=grad_neg_ln_like,
method="L-BFGS-B",
tol=1e-5)
# Update the kernel and print the final log-likelihood.
gp.set_parameter_vector(results.x)
except:
gp, results = kernel1(data)
return gp, results
offset = np.zeros(len(t))
idx = t < 57161
offset[idx] = self.d_harps1
offset[~idx] = self.d_harps2
return rv1 + rv2 + offset
#==============================================================================
# GP
#==============================================================================
from george import kernels
if star == 'HD117618':
k1 = kernels.ExpSine2Kernel(gamma=1,
log_period=np.log(100),
bounds=dict(gamma=(0, 100),
log_period=(0, 10)))
k2 = np.std(y) * kernels.ExpSquaredKernel(100)
kernel = k1 * k2
truth = dict(P1=0.25,
tau1=0.1,
k1=np.std(y) / 100,
w1=0.,
e1=0.4,
P2=3.1,
tau2=0.1,
k2=np.std(y) / 100,
w2=0.,
e2=0.4,
d_aat=0.,
d_harps1=0.,
offset = np.zeros(len(t))
idx = t < 57161
offset[idx] = self.d_harps1
offset[~idx] = self.d_harps2
return rv1 + rv2 + offset
#==============================================================================
# GP
#==============================================================================
from george import kernels
if star == 'HD117618':
k1 = kernels.ExpSine2Kernel(gamma=1,
log_period=np.log(100),
bounds=dict(gamma=(-1, 100),
log_period=(0, 10)))
k2 = np.std(y) * kernels.ExpSquaredKernel(1.)
k3 = 0.66**2 * kernels.RationalQuadraticKernel(log_alpha=np.log(0.78),
metric=1.2**2)
kernel = k1 * k2 + k3
truth = dict(P1=0.25,
tau1=0.1,
k1=np.std(y) / 100,
w1=0.,
e1=0.4,
P2=3.1,
tau2=0.1,
k2=np.std(y) / 100,
w2=0.,
e2=0.4,
def gaussian_process_smooth(self,
per=None,
minobs=10,
phase_offset=None,
recompute=False,
scalemin=None,
scalemax=None):
"""
per = cheaty hackish thing to get a gaussian process with some
continuity at the end points
minobs = mininum number of observations in each filter before we fit
"""
outgp = getattr(self, 'outgp', None)
if outgp is not None:
if not recompute:
return outgp
else:
outgp = {}
# note that we explicitly ask for non-smoothed and non-gpr lc here, since we're trying to compute the gpr
outlc = self.get_lc(recompute=False,
per=per,
smoothed=False,
gpr=False,
phase_offset=phase_offset)
for i, pb in enumerate(outlc):
thislc = outlc.get(pb)
thisphase, thisFlux, thisFluxErr = thislc
nobs = len(thisphase)
# if we do not have enough observations in this passband, skip it
if nobs < minobs:
continue
# TODO : REVISIT KERNEL CHOICE
if per == 1:
# periodic
kernel = kernels.ConstantKernel(1.) * kernels.ExpSine2Kernel(
1.0, 0.0)
elif per == 2:
# quasiperiodic
kernel = kernels.ConstantKernel(1.) * kernels.ExpSquaredKernel(
100.) * kernels.ExpSine2Kernel(1.0, 0.0)
else:
# non-periodic
kernel = kernels.ConstantKernel(1.) * kernels.ExpSquaredKernel(
1.)
gp = george.GP(kernel,
mean=thisFlux.mean(),
fit_mean=True,
fit_white_noise=True,
white_noise=np.log(thisFluxErr.mean()**2.))
# define the objective function
def nll(p):
gp.set_parameter_vector(p)
ll = gp.lnlikelihood(thisFlux, quiet=True)
return -ll if np.isfinite(ll) else 1e25
# define the gradient of the objective function.
def grad_nll(p):
gp.set_parameter_vector(p)
return -gp.grad_lnlikelihood(thisFlux, quiet=True)
# pre-compute the kernel
gp.compute(thisphase, thisFluxErr)
p0 = gp.get_parameter_vector()
max_white_noise = np.log((3. * np.median(thisFluxErr))**2.)
min_white_noise = np.log((0.3 * np.median(thisFluxErr))**2)
# coarse optimization with scipy.optimize
# TODO : almost anything is better than scipy.optimize
if per == 1:
# mean, white_noise, amplitude, gamma, FIXED_period
results = op.minimize(nll, p0, jac=grad_nll, bounds=[(None, None), (min_white_noise, max_white_noise),\
(None, None), (None, None), (0.,0.)])
elif per == 2:
# mean white_noise, amplitude, variation_timescale, gamma, FIXED_period
results = op.minimize(nll, p0, jac=grad_nll, bounds=[(None, None), (min_white_noise, max_white_noise),\
(None, None), (scalemin, scalemax), (None, None), (0.,0.)])
else:
# mean white_noise, amplitude, variation_timescale
results = op.minimize(nll, p0, jac=grad_nll, bounds=[(None, None), (min_white_noise, max_white_noise),\
(None, None), (scalemin,scalemax)])
gp.set_parameter_vector(results.x)
# george is a little different than sklearn in that the prediction stage needs the input data
outgp[pb] = (gp, thisphase, thisFlux, thisFluxErr)
self.outgp = outgp
return outgp
# Plot the data using the star.plot function
# plt.plot(RV_ALL[:,0], y, '-')
plt.plot(t, y, '-')
plt.ylabel('RV' r"$[m/s]$")
plt.xlabel('t')
plt.title('Simulated RV')
plt.show()
#==============================================================================
# GP Modelling
#==============================================================================
from george.modeling import Model
from george import kernels
k1 = kernels.ExpSine2Kernel(gamma=1, log_period=np.log(415.4))
k2 = np.var(y) * kernels.ExpSquaredKernel(1)
kernel = k1 * k2
gp = george.GP(kernel,
mean=Model(**truth),
white_noise=np.log(1),
fit_white_noise=True)
# gp = george.GP(kernel, mean=Model(**truth))
gp.compute(RV_ALL[:, 0], RV_ALL[:, 2])
def lnprob2(p):
# Set the parameter values to the given vector
gp.set_parameter_vector(p)
t = t_bin y = y_bin yerr = err_bin else: t = BJD y = Jitter yerr = err ################ # Optimization # ################ from george import kernels # kernel1 = 2.0 * kernels.Matern32Kernel(5) kernel2 = 10 * kernels.ExpSine2Kernel(gamma=10.0, log_period=np.log(8)) kernel2 *= kernels.ExpSquaredKernel(5) # kernel = kernel1 + kernel2 kernel = kernel2 import george # gp = george.GP(kernel) # y_test = gp.sample(BJD) # gp.compute(BJD, yerr) # plt.errorbar(t, y_test, yerr=yerr, fmt=".k", capsize=0) # plt.show() #gp = george.GP(kernel, mean=np.mean(y), fit_mean=True, # white_noise=np.log(0.19**2), fit_white_noise=True) gp = george.GP(kernel, mean=np.mean(y), fit_mean=True)
import numpy as np
import george
from george import kernels
from celerite import terms
__all__ = ["george_kernels", "george_solvers",
"celerite_terms",
"setup_george_kernel", "setup_celerite_terms"]
george_kernels = {
"Exp2": kernels.ExpSquaredKernel(10**2),
"Exp2ESin2": (kernels.ExpSquaredKernel(10**2) *
kernels.ExpSine2Kernel(2 / 1.3**2, 1.0)),
"ESin2": kernels.ExpSine2Kernel(2 / 1.3**2, 1.0),
"27dESin2": kernels.ExpSine2Kernel(2 / 1.3**2, 27.0 / 365.25),
"RatQ": kernels.RationalQuadraticKernel(0.8, 0.1**2),
"Mat32": kernels.Matern32Kernel((0.5)**2),
"Exp": kernels.ExpKernel((0.5)**2),
# "W": kernels.WhiteKernel, # deprecated, delegated to `white_noise`
"B": kernels.ConstantKernel,
}
george_solvers = {
"basic": george.BasicSolver,
"HODLR": george.HODLRSolver,
}
celerite_terms = {
"N": terms.Term(),
plt.errorbar(t, y, yerr=yerr, fmt=".k", capsize=0)
plt.ylabel(r"$y$ [m/s]")
plt.xlabel(r"$t$ [days]")
plt.ylim((-8, 12))
plt.title("Simulated observed data")
plt.show()
#==============================================================================
# Modelling
#==============================================================================
import george
george.__version__
from george import kernels
k1 = kernels.ExpSine2Kernel(gamma=6, log_period=np.log(25))
k2 = np.var(y) * kernels.ExpSquaredKernel(1)
kernel = k1 * k2
# mean: An object (following the modeling protocol) that specifies the mean function of the GP.
#gp = george.GP(k, mean=Model(amp=4.4, P=7.6, phase=0))
gp = george.GP(kernel,
mean=Model(**truth),
white_noise=np.log(1),
fit_white_noise=True)
gp.compute(t, yerr)
def lnprob2(p):
# Set the parameter values to the given vector
gp.set_parameter_vector(p)
# Compute the logarithm of the marginalized likelihood of a set of observations under the Gaussian process model.
t = t_bin
y = y_bin
yerr = err_bin
else:
t = BJD
y = Jitter
yerr = RV_noise
#==============================================================================
# Gaussian Processes
#==============================================================================
import george
from george import kernels
# kernel1 = kernels.Matern32Kernel(5)
kernel2 = np.var(y) * kernels.ExpSine2Kernel(gamma=2.,
log_period=np.log(23.81 / 3))
kernel2 *= kernels.ExpSquaredKernel(5)
# kernel = kernel1 + kernel2
kernel = kernel2
# kernel.freeze_parameter("k1:k2:log_period")
if 0: # generate a test GP time series
gp = george.GP(kernel)
y_test = gp.sample(BJD)
plt.errorbar(t, y_test, yerr=yerr, fmt=".k", capsize=0)
plt.show()
#gp = george.GP(kernel, mean=np.mean(y), fit_mean=True,
# white_noise=np.log(0.19**2), fit_white_noise=True)
# gp = george.GP(kernel, mean=np.mean(y), fit_mean=True)
gp = george.GP(kernel)
x = np.random.rand(100)
gp.compute(x, 1e-2)
kernels_to_test = [
kernels.ConstantKernel(log_constant=0.1),
kernels.ConstantKernel(log_constant=10.0, ndim=2),
kernels.ConstantKernel(log_constant=5.0, ndim=5),
kernels.DotProductKernel(),
kernels.DotProductKernel(ndim=2),
kernels.DotProductKernel(ndim=5, axes=0),
kernels.CosineKernel(log_period=1.0),
kernels.CosineKernel(log_period=0.5, ndim=2),
kernels.CosineKernel(log_period=0.5, ndim=2, axes=1),
kernels.CosineKernel(log_period=0.75, ndim=5, axes=[2, 3]),
kernels.ExpSine2Kernel(gamma=0.4, log_period=1.0),
kernels.ExpSine2Kernel(gamma=12., log_period=0.5, ndim=2),
kernels.ExpSine2Kernel(gamma=17., log_period=0.5, ndim=2, axes=1),
kernels.ExpSine2Kernel(gamma=13.7, log_period=-0.75, ndim=5, axes=[2, 3]),
kernels.ExpSine2Kernel(gamma=-0.7, log_period=0.75, ndim=5, axes=[2, 3]),
kernels.ExpSine2Kernel(gamma=-10, log_period=0.75),
kernels.LocalGaussianKernel(log_width=0.5, location=1.0),
kernels.LocalGaussianKernel(log_width=0.1, location=0.5, ndim=2),
kernels.LocalGaussianKernel(log_width=1.5, location=-0.5, ndim=2, axes=1),
kernels.LocalGaussianKernel(log_width=2.0,
location=0.75,
ndim=5,
axes=[2, 3]),
kernels.LinearKernel(order=0, log_gamma2=0.0),
kernels.LinearKernel(order=2, log_gamma2=0.0),
kernels.LinearKernel(order=2, log_gamma2=0.0),
#==============================================================================
# GP
#==============================================================================
t = t[idx]
y = XY[idx]
yerr = yerr[idx]
from george import kernels
# k1 = 1**2 * kernels.ExpSquaredKernel(metric=10**2)
# k2 = 1**2 * kernels.ExpSquaredKernel(80**2) * kernels.ExpSine2Kernel(gamma=8, log_period=np.log(36.2))
# boundary doesn't seem to take effect
k2 = 1**2 * kernels.ExpSquaredKernel(80**2) * kernels.ExpSine2Kernel(
gamma=11,
log_period=np.log(36.2),
bounds=dict(gamma=(-3, 30),
log_period=(np.log(36.2 - 5), np.log(36.2 + 6))))
# k3 = 0.66**2 * kernels.RationalQuadraticKernel(log_alpha=np.log(0.78), metric=1.2**2)
# k4 = 1**2 * kernels.ExpSquaredKernel(40**2)
# kernel = k1 + k2 + k3 + k4
kernel = k2
import george
# gp = george.GP(kernel, mean=np.mean(y), fit_mean=True,
# white_noise=np.log(0.19**2), fit_white_noise=True)
gp = george.GP(kernel, mean=np.mean(y), fit_mean=True)
# gp.freeze_parameter('kernel:k2:log_period')
# gp.freeze_parameter('kernel:k2:gamma')
#==============================================================================
'''
mu = 0.07 # mean of the Normal prior
sigma = 0.004 # standard deviation of the Normal prior
prob_r = np.log(gaussian(r - mu, sigma))
mu = 2456915.70
sigma = 0.005
prob_t0 = np.log(gaussian(T0 - mu, sigma))
prob = prob_t0
'''
return 0
K = {'Constant': 2. ** 2,
'ExpSquaredKernel': kernels.ExpSquaredKernel(metric=1.**2),
'ExpSine2Kernel': kernels.ExpSine2Kernel(gamma=1.0, log_period=1.0),
'Matern32Kernel': kernels.Matern32Kernel(2.)}
def neo_init_george(kernels):
k_out = K['Constant']
for func in kernels[0]:
k_out *= K[func]
for i in range(len(kernels)):
if i == 0:
pass
else:
k = K['Constant']
for func in kernels[i]:
k *= K[func]
k_out += k
gp = george.GP(k_out)
if 0: # plot the jitter time series
plt.errorbar(t, y, yerr=yerr, fmt="o", capsize=0, alpha=0.5)
plt.ylabel('RV [m/s]')
plt.xlabel('time [day]')
plt.show()
#==============================================================================
# GP
#==============================================================================
from george import kernels
k1 = 10 * kernels.ExpSquaredKernel(metric=10**2)
# k2 = 1**2 * kernels.ExpSquaredKernel(90**2) * kernels.ExpSine2Kernel(gamma=1, log_period=1.9)
k2 = 20 * kernels.ExpSquaredKernel(200**2) * kernels.ExpSine2Kernel(
gamma=2,
log_period=np.log(17),
bounds=dict(gamma=(-3, 30), log_period=(np.log(17 - 5), np.log(17 + 5))))
k3 = 20 * kernels.RationalQuadraticKernel(log_alpha=np.log(100), metric=120**2)
k4 = 10 * kernels.ExpSquaredKernel(1000**2)
# kernel = k1 + k2 + k3 + k4
kernel = k3 + k2 + k4
import george
#gp = george.GP(kernel, mean=np.mean(y), fit_mean=True,
# white_noise=np.log(0.19**2), fit_white_noise=True)
gp = george.GP(kernel, mean=np.mean(y), fit_mean=True)
gp.compute(t, yerr)
#==============================================================================
# Optimization
#==============================================================================
f2 = 2*np.arctan( np.sqrt((1+self.e2)/(1-self.e2))*np.tan(e_anom2*.5) )
rv2 = 100*self.k2*(np.cos(f2 + self.w2) + self.e2*np.cos(self.w2))
offset = np.zeros(len(t))
idx = t < 57161
offset[idx] = self.offset1
offset[~idx]= self.offset2
return rv1 + rv2 + offset
#==============================================================================
# GP
#==============================================================================
from george import kernels
k1 = kernels.ExpSine2Kernel(gamma = 1, log_period = np.log(3200),
bounds=dict(gamma=(-3,1), log_period=(0,10)))
k2 = kernels.ConstantKernel(log_constant=np.log(1.), bounds=dict(log_constant=(-5,5))) * kernels.ExpSquaredKernel(1.)
kernel = k1 * k2
truth = dict(P1=8., tau1=1., k1=np.std(y)/100, w1=0., e1=0.4,
P2=100, tau2=1., k2=np.std(y)/100, w2=0., e2=0.4, offset1=0., offset2=0.)
kwargs = dict(**truth)
kwargs["bounds"] = dict(P1=(7.5,8.5), k1=(0,0.1), w1=(-2*np.pi,2*np.pi), e1=(0,0.9),
tau2=(-50,50), k2=(0,0.2), w2=(-2*np.pi,2*np.pi), e2=(0,0.9))
mean_model = Model(**kwargs)
gp = george.GP(kernel, mean=mean_model, fit_mean=True)
# gp = george.GP(kernel, mean=mean_model, fit_mean=True, white_noise=np.log(0.5**2), fit_white_noise=True)
gp.compute(t, yerr)
def lnprob2(p):
gp.set_parameter_vector(p)
# Plot the data.
fig = pl.figure(figsize=(6, 3.5))
ax = fig.add_subplot(111)
ax.plot(t, y, ".k", ms=2)
ax.set_xlim(min(t), 1999)
ax.set_ylim(min(y), 369)
ax.set_xlabel("year")
ax.set_ylabel("CO$_2$ in ppm")
fig.subplots_adjust(left=0.15, bottom=0.2, right=0.99, top=0.95)
fig.savefig("../_static/hyper/data.png", dpi=150)
# Initialize the kernel.
k1 = 66.0**2 * kernels.ExpSquaredKernel(67.0**2)
k2 = 2.4**2 * kernels.ExpSquaredKernel(90**2) \
* kernels.ExpSine2Kernel(2.0 / 1.3**2, 1.0)
k3 = 0.66**2 * kernels.RationalQuadraticKernel(0.78, 1.2**2)
k4 = 0.18**2 * kernels.ExpSquaredKernel(1.6**2) + kernels.WhiteKernel(0.19)
kernel = k1 + k2 + k3 + k4
# Set up the Gaussian process and maximize the marginalized likelihood.
gp = george.GP(kernel, mean=np.mean(y))
# Define the objective function (negative log-likelihood in this case).
def nll(p):
# Update the kernel parameters and compute the likelihood.
gp.kernel[:] = p
ll = gp.lnlikelihood(y, quiet=True)
# The scipy optimizer doesn't play well with infinities.
return -ll if np.isfinite(ll) else 1e25
offset = np.zeros(len(t))
idx = t < 57300
offset[idx] = self.offset1
offset[~idx] = self.offset2
return rv1 + rv2 + offset
#==============================================================================
# GP
#==============================================================================
from george import kernels
k1 = 1**2 * kernels.ExpSquaredKernel(metric=10**2)
k2 = 1**2 * kernels.ExpSquaredKernel(90**2) * kernels.ExpSine2Kernel(
gamma=1, log_period=1.8)
k3 = 0.66**2 * kernels.RationalQuadraticKernel(log_alpha=np.log(0.78),
metric=1.2**2)
kernel = k1 + k2 + k3
truth = dict(P1=8.,
tau1=1.,
k1=np.std(y) / 100,
w1=0.,
e1=0.4,
P2=100,
tau2=1.,
k2=np.std(y) / 100,
w2=0.,
e2=0.4,
offset1=0.,
def predict_Q_data(input_dir_2, filename4="training.txt"):
nu, lam_squared, stokesQ, stokesU = read_data(input_dir_2, filename4)
nu_R = nu
stokesQ_R = stokesQ
lam_R = lam_squared
# Squared exponential kernel
k1 = 0.3**2 * kernels.ExpSquaredKernel(0.02**2)
# periodic covariance kernel with exponential component toallow decay away from periodicity
k2 = 0.6**2 * kernels.ExpSquaredKernel(0.5**2) * kernels.ExpSine2Kernel(
gamma=2 / 2.5**2, log_period=0.0)
###vary gamma value to widen the funnel
# rational quadratic kernel for medium term irregularities.
k3 = 0.3**2 * kernels.RationalQuadraticKernel(log_alpha=np.log(0.1),
metric=0.1**2)
# noise kernel: includes correlated noise & uncorrelated noise
k4 = 0.18**2 * kernels.ExpSquaredKernel(1.6**2)
kernel = k1 + k2 #+k3 + k4
gp = george.GP(kernel,
mean=np.mean(stokesQ),
fit_mean=True,
white_noise=np.log(0.02**2),
fit_white_noise=True)
#gp = george.GP(kernel)
gp.compute(lam_R)
# range of times for prediction:
x = np.linspace(np.min(lam_squared) - 0.01, np.max(lam_squared),
1024) #extend to smaller wavelengths
# calculate expectation and variance at each point:
mu_initial, cov = gp.predict(stokesQ_R, x)
#mu, cov = gp.predict(stokesQ_2Gz, x, return_var = True)
std_initial = np.sqrt(np.diag(cov))
print("Initial prediction of missing Q values: \n {}".format(mu_initial))
# Define the objective function (negative log-likelihood in this case).
def nll(p):
gp.set_parameter_vector(p)
ll = gp.log_likelihood(stokesQ_R, quiet=True)
return -ll if np.isfinite(ll) else 1e25
# And the gradient of the objective function.
def grad_nll(p):
gp.set_parameter_vector(p)
return -gp.grad_log_likelihood(stokesQ_R, quiet=True)
gp.compute(lam_R)
p0 = gp.get_parameter_vector()
results = op.minimize(nll, p0, jac=grad_nll, method="L-BFGS-B")
# run optimization:
#results = op.minimize(nll, p0, jac=grad_nll)
gp.set_parameter_vector(results.x)
x = np.linspace(np.min(lam_squared) - 0.01, np.max(lam_squared),
1024) #extend to smaller wavelengths
mu_optimized, cov = gp.predict(stokesQ_R, x)
std_optimized = np.sqrt(np.diag(cov))
print("Final prediction of missing Q values: \n {}".format(mu_optimized))
return nu_R, lam_R, stokesQ_R, mu_initial, std_initial, mu_optimized, std_optimized
#Load data
data = co2.load_pandas().data
t = 2000 + (np.array(data.index.to_julian_date()) - 2451545.0) / 365.25
y = np.array(data.co2)
m = np.isfinite(t) & np.isfinite(y) & (t < 1996)
t, y = t[m][::4], y[m][::4]
plt.plot(t, y, ".k")
plt.xlim(t.min(), t.max())
plt.xlabel("year")
plt.ylabel("CO$_2$ in ppm")
#Load kernels
k1 = 66**2 * kernels.ExpSquaredKernel(metric=67**2)
k2 = 2.4**2 * kernels.ExpSquaredKernel(90**2) * kernels.ExpSine2Kernel(
gamma=2 / 1.3**2, log_period=0.0)
k3 = 0.66**2 * kernels.RationalQuadraticKernel(log_alpha=np.log(0.78),
metric=1.2**2)
k4 = 0.18**2 * kernels.ExpSquaredKernel(1.6**2)
kernel = k1 + k2 + k3 + k4
gp = george.GP(kernel,
mean=np.mean(y),
fit_mean=True,
white_noise=np.log(0.19**2),
fit_white_noise=True)
gp.compute(t)
print(gp.log_likelihood(y))
print(gp.grad_log_likelihood(y))
import scipy.optimize as op
strLC = KeplerLightCurveFile.from_archive(206208968)
strPDC = strLC.PDCSAP_FLUX.remove_outliers()
y = strPDC.flux[:300]
x = strPDC.time[:300]
y = (y / np.median(y)) - 1 # sets the function to begin at 0
x = x[np.isfinite(y)]
y = y[np.isfinite(y)] # removes NaN values
pl.plot(x, y)
#pl.show()
print(median_absolute_deviation(y))
print(np.var(y))
kernel = np.var(y) * kernels.ExpSquaredKernel(0.5) * kernels.ExpSine2Kernel(
log_period=0.5, gamma=1)
gp = george.GP(kernel)
gp.compute(x, y)
result = minimize(neg_ln_like, gp.get_parameter_vector(), jac=grad_neg_ln_like)
gp.set_parameter_vector(result.x)
#def plotModel ():
pred_mean, pred_var = gp.predict(y, x, return_var=True)
pred_std = np.sqrt(pred_var)
#pl.errorbar(x,y, yerr=np.var(y) ** 0.5/10, fmt='k.')
pl.plot(x, pred_mean, '--r')
pl.plot(x, y, lw=0.1, color='blue')
pl.ylabel('Relative Flux')
pl.xlabel('Time - BKJDC')
pl.show()
