def plot_posterior_nodes(model, param_nodes, bins=100, lb=None, ub=None):
sns.set_style('white')
sns.despine()
#title='Generic Title'
if param_nodes=='z':
nodes=get_nodes(model, 'z')
xlabel='Mean Starting-Point' + r'$\/(\mu_{z})$'
lb=.52
ub=.67
elif param_nodes=='vf':
nodes=get_nodes(model, 'vf')
xlabel='Mean Face Drift-Rate' + r'$\/(\mu_{vF})$'
lb=.35
ub=1.2
elif param_nodes=='vh':
nodes=get_nodes(model, 'vh')
xlabel='Mean House Drift-Rate' + r'$\/(\mu_{vH})$'
lb=-1.15
ub=-.25
else:
print "Must provide argument: 'z', 'vf', or 'vh'"
fig=plt.figure()
fig.subplots_adjust(top=0.95, wspace=0.12, left=0.12, right=0.88, bottom=0.16)
sns.despine()
#fig.suptitle(title, fontsize=20)
#fig.suptitle(title, fontsize=40)
if lb is None:
lb = min([min(node.trace()[:]) for node in nodes])
if ub is None:
ub = max([max(node.trace()[:]) for node in nodes])
x_data = np.linspace(lb, ub, 600)
#colors=['Green', 'LimeGreen', 'Black', 'Cyan', 'Blue']
colors=['Red','Magenta', 'Black', 'Cyan', 'Blue']
color_i=0
for node in nodes:
trace = node.trace()[:]
#hist = interpolate_trace(x_data, trace, range=(trace.min(), trace.max()), bins=bins)
hist = interpolate_trace(x_data, trace, range=(lb, ub), bins=bins)
plt.plot(x_data, hist, label=node.__name__, lw=2., color=colors[color_i])
plt.fill_between(x_data, hist, 0, label=node.__name__, color=colors[color_i], alpha=0.3)
ax=plt.gca()
ax.set_xlim(lb, ub)
plt.setp(ax.get_yticklabels(), visible=False)
sns.despine()
ax.set_ylabel('Probability Mass', fontsize=22, labelpad=12)
ax.set_xlabel(xlabel, fontsize=22, labelpad=13)
plt.setp(ax.get_xticklabels(), fontsize=18)
plt.locator_params(axis='x', nbins=10)
color_i+=1
#leg = plt.legend(loc='best', fancybox=True)
#leg.get_frame().set_alpha(0.5)
plt.ylim(ymin=0)
plt.savefig(str(param_nodes)+'_posterior_nodes.png', dpi=600)
def savage_dickey(pos, post_trace, range=(-.3,.3), bins=40, prior_trace=None, prior_y=None):
"""Calculate Savage-Dickey density ratio test, see Wagenmakers et
al. 2010 at http://dx.doi.org/10.1016/j.cogpsych.2009.12.001
:Arguments:
pos : float
position at which to calculate the savage dickey ratio at (i.e. the spec hypothesis you want to test)
post_trace : numpy.array
trace of the posterior distribution
:Optional:
prior_trace : numpy.array
trace of the prior distribution
prior_y : numpy.array
prior density pos
range : (int,int)
Range over which to interpolate and plot
bins : int
Over how many bins to compute the histogram over
:Note: Supply either prior_trace or prior_y.
"""
x = np.linspace(range[0], range[1], bins)
if prior_trace is not None:
# Prior is provided as a trace -> histogram + interpolate
prior_pos = interpolate_trace(pos, prior_trace, range=range, bins=bins)
elif prior_y is not None:
# Prior is provided as a density for each point -> interpolate to retrieve positional density
import scipy.interpolate
prior_pos = prior_y #scipy.interpolate.InterpolatedUnivariateSpline(x, prior_y)(pos)
else:
assert ValueError, "Supply either prior_trace or prior_y keyword arguments"
# Histogram and interpolate posterior trace at SD position
posterior_pos = interpolate_trace(pos, post_trace, range=range, bins=bins)
# Calculate Savage-Dickey density ratio at pos
sav_dick = prior_pos / posterior_pos
return sav_dick
def plot_posteriors(nodes, bins=50):
from kabuki.utils import interpolate_trace
figure()
lb = min([min(node.trace()[:]) for node in nodes])
ub = max([max(node.trace()[:]) for node in nodes])
x_data = np.linspace(lb, ub, 300)
for node in nodes:
trace = node.trace()[:]
#hist = interpolate_trace(x_data, trace, range=(trace.min(), trace.max()), bins=bins)
hist = interpolate_trace(x_data, trace, range=(lb, ub), bins=bins)
plt.plot(x_data, hist, label=node.__name__, lw=2.)
leg = plt.legend(loc='best', fancybox=True)
leg.get_frame().set_alpha(0.5)
def plot_posterior_nodes(nodes, bins=50):
"""Plot interpolated posterior of a list of nodes.
:Arguments:
nodes : list of pymc.Node's
List of pymc.Node's to plot the posterior of
bins : int (default=50)
How many bins to use for computing the histogram.
"""
from kabuki.utils import interpolate_trace
figure()
lb = min([min(node.trace()[:]) for node in nodes])
ub = max([max(node.trace()[:]) for node in nodes])
x_data = np.linspace(lb, ub, 300)
for node in nodes:
trace = node.trace()[:]
#hist = interpolate_trace(x_data, trace, range=(trace.min(), trace.max()), bins=bins)
hist = interpolate_trace(x_data, trace, range=(lb, ub), bins=bins)
plt.plot(x_data, hist, label=node.__name__, lw=2.)
leg = plt.legend(loc='best', fancybox=True)
leg.get_frame().set_alpha(0.5)