kalman.LinearGauss.__init__(self, sigmaX=sigmaX, sigmaY=sigmaY, rho=rho,
sigma0=sigma0)
data = np.loadtxt('./simulated_linGauss_T100_varX1_varY.04_rho.9.txt')
niter = 10 ** 5
burnin = int(niter/ 10)
algos = OrderedDict()
# PMMH algorithms
scales = list(reversed([0.025, 0.05, 0.1, 0.15, 0.2, 0.3, 0.4, 0.7, 1.]))
for scale in scales:
key = 'scale=%g' % scale
algos[key] = mcmc.PMMH(ssm_cls=ReparamLinGauss, prior=prior,
data=data, Nx=100, niter=niter,
adaptive=False, rw_cov=scale ** 2 * np.eye(3),
verbose=10, scale=scale)
# Run the algorithms
####################
for alg_name, alg in algos.items():
print('\nRunning ' + alg_name)
alg.run()
print('CPU time: %.2f min' % (alg.cpu_time / 60))
print('mean squared jumping distance: %f' % msjd(alg.chain.theta[burnin:]))
# Plots
#######
savefigs = False
plt.style.use('ggplot')
# data was simulated as follows:
# _, data = ReparamLinGauss(varX=1., varY=(0.2)**2, rho=.9).simulate(100)
data = np.loadtxt('./simulated_linGauss_T100_varX1_varY.04_rho.9.txt')
niter = 10**3
algos = OrderedDict()
rw_cov = (0.15)**2 * np.eye(3)
# PMMH algorithms
Nxs = list(range(100, 1100, 100)) # needs list for Python 3
for Nx in Nxs:
algos[Nx] = mcmc.PMMH(ssm_cls=ReparamLinGauss,
prior=prior,
data=data,
Nx=Nx,
niter=niter,
adaptive=False,
rw_cov=rw_cov,
verbose=10)
# Run the algorithms
####################
for Nx, alg in algos.items():
print('Nx = %i' % Nx)
alg.run()
print('CPU time: %.2f min' % (alg.cpu_time / 60))
# PLOTS
#######
savefigs = True # False if you don't want to save plots as pdfs
#sb.boxplot(x=[r['output'].logLt for r in results], y=[r['qmc'] for r in results])
smooth_trajectories = pf.hist.backward_sampling(10)
plt.figure()
plt.plot(smooth_trajectories)
prior_dict = {
'mu': dists.Normal(),
'sigma': dists.Gamma(a=1., b=1.),
'rho': dists.Beta(9., 1.)
}
my_prior = dists.StructDist(prior_dict)
from particles import mcmc # where the MCMC algorithms (PMMH, Particle Gibbs, etc) live
pmmh = mcmc.PMMH(ssm_cls=StochVol,
prior=my_prior,
data=data,
Nx=50,
niter=1000)
pmmh.run() # Warning: takes a few seconds
plt.figure()
# plot the marginals
burnin = 100 # discard the 100 first iterations
for i, param in enumerate(prior_dict.keys()):
plt.subplot(2, 2, i + 1)
sb.distplot(pmmh.chain.theta[param][burnin:], 40)
plt.title(param)
plt.show()