msg = "model evidence: " + str(D['lnZ']) + \
" +/- " + str(D['dlnZ'])
cc.WP(msg, D['wrt_file'])
cp.plot_hist(flattrace,
D['pname'],
D['pname_plt'],
param_true=D['param_true'],
pltname=D['outname'])
cc.coef_summary(flattrace, D['pname'], D['outname'])
bounds = ((.25, 2.1), (-.8, 2.2), (0, .6))
cp.plot_cov(flattrace,
D['pname_plt'],
param_true=D['param_true'],
bounds=bounds,
figsize=[5.5, 5.5],
pltname=D['outname'])
cp.plot_prediction(flattrace,
D['name_list'],
D['Tt'],
D['At'],
D['It'],
feval,
D,
colorL=['k', 'r'],
param_true=D['param_true'],
ylim=(1.1, 2.2),
pltname=D['outname'])
WP(msg, D['wrt_file'])
msg = "model evidence: " + str(D['lnZ']) + \
" +/- " + str(D['dlnZ'])
cc.WP(msg, D['wrt_file'])
cc.coef_summary(flattrace, D['pname'], D['wrt_file'])
nxtprior = np.zeros((2, D['nparam']))
nxtprior[0, :] = np.mean(flattrace, 0) - 5*np.std(flattrace, 0)
nxtprior[1, :] = 10*np.std(flattrace, 0)
np.savetxt(D['outname'] + '_prior.csv', nxtprior)
cp.plot_hist(flattrace, D['pname'], D['pname_plt'], pltname=D['outname'])
cp.plot_cov(flattrace, D['pname_plt'], pltname=D['outname'],
tight_layout=False)
"""configure model prediction plots for Cp, H, S and G"""
name_list_l = [D['name_list_Cp'], D['name_list_H'], None, None]
Tt_l = [D['Tt_Cp'], D['Tt_H'], None, None]
At_l = [D['At_Cp'], D['At_H'], None, None]
Etr_l = [D['Etr_Cp'], D['Etr_H'], None, None]
It_l = [D['It_Cp'], D['It_H'], None, None]
pltper = [1, 1, 0, 0]
xlim = [[(1800, 2600)], [(1800, 2600)], None, None]
ylim = [[(24, 40)], [(47000, 85000)], None, None]
on = D['outname']
pltname = [[on + 'Cp', on + 'Cp_close', on + 'Cp_vclose'],
[on + 'H'],
[on + 'S'],
[on + 'G']]
msg = "model evidence: " + str(D['lnZ']) + \
" +/- " + str(D['dlnZ'])
cc.WP(msg, D['wrt_file'])
cp.plot_hist(flattrace,
D['pname'],
D['pname_plt'],
param_true=D['param_true'],
pltname=D['outname'])
cc.coef_summary(flattrace, D['pname'], D['outname'])
cp.plot_cov(flattrace[:, :3],
D['pname_plt'][3:],
param_true=D['param_true'][3:],
figsize=[5.5, 5.5],
sciform=True,
pltname=D['outname'] + '_Cp')
cp.plot_cov(flattrace[:, 3:],
D['pname_plt'][3:],
param_true=D['param_true'][3:],
figsize=[5.5, 5.5],
sciform=True,
pltname=D['outname'] + '_H')
cp.plot_prediction(flattrace,
D['name_list'],
D['Tt_Cp'],
D['At_Cp'],
D['It_Cp'],
msg = "model evidence: " + str(D['lnZ']) + \
" +/- " + str(D['dlnZ'])
cc.WP(msg, D['wrt_file'])
cp.plot_hist(flattrace,
D['pname'],
D['pname_plt'],
param_true=D['param_true'],
pltname=D['outname'])
cc.coef_summary(flattrace, D['pname'], D['outname'])
cp.plot_cov(flattrace,
D['pname_plt'],
param_true=D['param_true'],
figsize=[5.5, 5.5],
pltname=D['outname'],
sciform=True)
# rescale the data errors by the means of the hyperparameters
hyp_m = np.mean(flattrace[:, D['order'] + 1:], 0)
hyp_m_V = np.zeros(D['Et'].shape)
for ii in range(len(hyp_m)):
hyp_m_V[D['It'] == ii] = hyp_m[ii]
cp.plot_prediction(flattrace,
D['name_list'],
D['Tt'],
D['At'],
D['It'],
feval,