def compareTimes(table_names = ['FinalEstimate_4x100_N=100','FinalEstimate_4x100_N=1000']):
regime_labels = {'subT': 'Sub-threshold',
'superT': 'Supra-threshold',
'superSin': 'Supra-sinusoidal',
'crit': 'Critical'}
for table_name, N in zip(table_names,
[100, 1000]):
analyzer = DataAnalyzer(table_name);
print table_name
for regime in analyzer.getAllRegimes():
print regime_labels[regime]
for mthd in [ 'Fortet', 'FP']:
walltimes = analyzer.getWallTimes(regime, mthd)
mu, sigma = mean(walltimes), std(walltimes)
print '& %.2f'%mu, r'$\pm$ %.2f'%sigma
print r'\\'
def boxPlotBox():
table_file_name = 'FinalEstimate_4x100_N=100'
Methods = ['Initializer', 'FP', 'Fortet']
regime = 'crit'
analyzer = DataAnalyzer(table_file_name)
mpl.rcParams['figure.figsize'] = 17, 6
method_labels = ['Initializer', 'Numerical FP', 'Fortet']
xlabel_font_size = 12
abg_true = analyzer.getTrueParamValues(regime)
print regime, abg_true
figure()
lgs = GridSpec(1, 3)
abgs = {}
for mthd in Methods:
abgs[mthd] = array(analyzer.getAllEstimates(regime, mthd))
# method_xpos = dict.fromkeys(Methods, [1,2,3])
param_tags = [r' \alpha ', r' \beta ', r' \gamma ']
for params_idx, param_tag in enumerate(param_tags):
print params_idx, param_tag
ax_low = subplot(lgs[params_idx]); hold(True)
# for mthd in Methods:
param_ests = [ abgs[mthd][:,params_idx] for mthd in Methods]
param_true = abg_true[params_idx]
boxplot(param_ests, positions = [1,3,5])
hlines(param_true, 0,5, linestyles='dashed')
title(r'$%s = %.2f$'%(param_tag, param_true))
ylabel(r'$%s$'%param_tag, fontsize = xlabel_font_size)
xtickNames = setp(ax_low, xticklabels = method_labels)
setp(xtickNames, rotation=60, fontsize=xlabel_font_size)
#
subplots_adjust(left=0.1, right=0.95, top=0.9, bottom=0.35,
wspace = 0.65, hspace = 0.2)
def postVisualizer():
N_phi = 20
print 'N_phi = ', N_phi
phi_norms = linspace(1/(2.*N_phi), 1. - 1/ (2.*N_phi), N_phi)
theta = 20
base_name = 'sinusoidal_spike_train_N=1000_critical_theta=%d'%theta
T_thresh = 64.
analyzer = DataAnalyzer('ThetaEstimate_4x100_N=1000')
sample_id = 32
regime_label = base_name + '%d'%theta
file_name = base_name + '_%d'%sample_id
print file_name
binnedTrain = BinnedSpikeTrain.initFromFile(file_name, phi_norms)
binnedTrain.pruneBins(None, N_thresh = 1, T_thresh=T_thresh)
regime_name = 'theta%d'%theta
abg_true = analyzer.getTrueParamValues(regime_name);
print abg_true
abg_fortet = analyzer.getEstimates(sample_id, regime_name, 'Fortet')[0]
print abg_fortet
visualizeData_vs_Fortet(abg_fortet, binnedTrain, theta,title_tag = 'Fortet: estimates', save_fig_name='theta20_Fortet_estimates')
visualizeData_vs_Fortet(abg_true, binnedTrain, theta,title_tag = 'Fortet: true', save_fig_name='theta20_Fortet_true')
abg_fp = analyzer.getEstimates(sample_id, regime_name, 'FP')[0]
print abg_fp
dx = .025; dt = FPMultiPhiSolver.calculate_dt(dx, abg_true, -1.0)
phis = binnedTrain.bins.keys();
S = FPMultiPhiSolver(binnedTrain.theta, phis,
dx, dt,
binnedTrain.getTf(), X_min = -1.0)
visualizeData_vs_FP(S, abg_fp, binnedTrain,title_tag = 'FP: estimates', save_fig_name='theta20_FP_estimates')
visualizeData_vs_FP(S, abg_true, binnedTrain,title_tag = 'FP: true', save_fig_name='theta20_FP_true')
def tabulateEstimates(table_file_names,
methods):
for table_file_name in table_file_names:
print table_file_name
analyzer = DataAnalyzer(table_file_name)
regimeNames = analyzer.getAllRegimes()
param_names = [r'\a',r'\b',r'\g']
for regime in regimeNames:
print r'\subfloat[%s]{\begin{tabular}{|c|ccc|} '%regime_labels[regime]
abg_true = analyzer.getTrueParamValues(regime)
print 'Parameter'
for mthd in methods:
print r'&', mthd
print r'\\ \hline'
for idx in xrange(3):
print r'$%s=%.2f$' %(param_names[idx],abg_true[idx])
for mthd in methods:
abg_est = array(analyzer.getAllEstimates(regime, mthd))
est = sort(abg_est[:, idx])
print r'& $%.2f : [%.2f, %.2f]$' %(mean(est), est[2], est[-3])
print r'\\'
print r' \end{tabular}}\\'
print 64*'#'
def crossAnalyze(table_file_name,
fig_id = '', Method1 = 'Fortet', Method2 = 'FP',
regimeNames = None):
analyzer = DataAnalyzer(table_file_name)
mpl.rcParams['figure.figsize'] = 17, 5*4
mpl.rcParams['figure.subplot.left'] = .15
mpl.rcParams['figure.subplot.right'] = .975
mpl.rcParams['figure.subplot.bottom'] = .05
mpl.rcParams['figure.subplot.top'] = .95
mpl.rcParams['figure.subplot.wspace'] = .25
mpl.rcParams['figure.subplot.hspace'] = .375
fig = figure();
if None==regimeNames:
regimeNames = analyzer.getAllRegimes()
plot_scale = 1.05;
regime_tags = {'superT':'Supra-Threshold', 'crit':'Critical',
'subT':'Sub-Threshold', 'superSin':'Super-Sinusoidal',
'theta1':r'$\Omega=1$',
'theta5':r'$\Omega=5$',
'theta10':r'$\Omega=10$',
'theta20':r'$\Omega=20$'}
xlabel_font_size = 32
for regime_idx,regime in enumerate(regimeNames):
abg_true = analyzer.getTrueParamValues(regime)
print regime, abg_true
abgs = {};
abgs[Method1] = array( analyzer.getAllEstimates(regime, Method1) )
abgs[Method2] = array( analyzer.getAllEstimates(regime, Method2) )
for param_idx, pname in zip(arange(3),
[r'\alpha', r'\beta', r'\gamma']):
axu_lim = amax( r_[ abgs[Method1][:, param_idx], abgs[Method2][:, param_idx]])
axl_lim = amin( r_[ abgs[Method1][:, param_idx], abgs[Method2][:, param_idx]])
plot_buffer = (plot_scale -1.)*(axu_lim - axl_lim);
axu_lim += plot_buffer
axl_lim -= plot_buffer
ax = subplot(len(regimeNames),3, regime_idx*3 + param_idx+1, aspect='equal')
plot(abgs[Method1][:, param_idx], abgs[Method2][:, param_idx],
linestyle='None',
color= 'k', marker= 'o', markersize = 7 )
if 0 == param_idx:
text(-.4, 0.5, regime_tags[regime],
horizontalalignment='center',
verticalalignment='center',
transform = ax.transAxes,
rotation='vertical',
size = xlabel_font_size)
# axu_lim, axl_lim = xlim()
hlines(abg_true[param_idx], axl_lim, axu_lim,
linestyles = 'dashed' )
# axu_lim, axl_lim = ylim()
vlines(abg_true[param_idx], axl_lim, axu_lim,
linestyles = 'dashed' )
ax.xaxis.set_major_locator(MaxNLocator(6))
ax.yaxis.set_major_locator(MaxNLocator(6))
xlim((axl_lim, axu_lim));
ylim((axl_lim, axu_lim));
xlabel(Method1,fontsize = xlabel_font_size);
ylabel(Method2,fontsize = xlabel_font_size);
title('$'+ pname + ' = %.2f$'%abg_true[param_idx], fontsize = 32)
t = add_inner_title(ax, chr(ord('A') +
regime_idx*3 + param_idx),
loc=2, size=dict(size=ABCD_LABEL_SIZE))
t.patch.set_ec("none")
t.patch.set_alpha(0.5)
if '' != fig_id:
file_name = os.path.join(FIGS_DIR, fig_id + '_cross_compare_joint.pdf')
print 'saving to ', file_name
savefig(file_name)
def postProcessThetaBoxPlots(table_file_name, fig_id = '',
Methods = ['Initializer', 'FP', 'Fortet'],
regimeNames = ['theta1', 'theta5', 'theta10', 'theta20'] ):
analyzer = DataAnalyzer(table_file_name)
if None==regimeNames:
regimeNames = analyzer.getAllRegimes()
mpl.rcParams['figure.figsize'] = 17, 12
figure()
subplots_adjust(left=0.1, right=0.975, top=0.9, bottom=0.15,
wspace = 0.3, hspace = 0.01)
lgs = GridSpec(4,3)
method_labels = ['Initializer', 'Numerical FP', 'Fortet']
xlabel_font_size = 24
sample_size_texts = ['N=100', 'N=1000']
yupper = 3*[-inf];
ylower = 3*[inf]
for params_idx in xrange(3):
for mthd in Methods:
for regime in regimeNames:
Omega = int(regime[5:])
ests = array(analyzer.getAllEstimates(regime, mthd))[:,params_idx]
if (2 == params_idx):
ests /= sqrt( 1. + Omega*Omega)
ylower[params_idx] = amin([ylower[params_idx], amin(ests)])
yupper[params_idx] = amax([yupper[params_idx], amax(ests)])
for regime_idx, regime in enumerate(regimeNames):
abg_true = analyzer.getTrueParamValues(regime)
Omega = int(regime[5:])
print regime, abg_true, Omega
abgs = {}
param_tags = [r' \alpha ', r' \beta ', r' \gamma / \sqrt{1+ \Omega}']
for mthd in Methods:
abgs[mthd] = array(analyzer.getAllEstimates(regime, mthd))
for params_idx, param_tag in enumerate(param_tags):
print params_idx, param_tag
subplot_idx = regime_idx*3 + params_idx;
ax = subplot(lgs[subplot_idx]); hold(True)
# for mthd in Methods:
param_ests = [ abgs[mthd][:,params_idx] for mthd in Methods]
param_true = abg_true[params_idx]
if params_idx == 2:
param_true /= sqrt(1. + Omega*Omega)
for idx in xrange(len(param_ests)):
param_ests[idx] /= sqrt(1. + Omega*Omega)
boxplot(param_ests, positions = [1,3,5])
hlines(param_true, 0,5, linestyles='dashed')
ylabel(r'$\hat{%s}$'%param_tag, fontsize = xlabel_font_size)
if (regime_idx ==0):
title(r'$%s = %.1f$'%(param_tag, param_true), fontsize = 32)
if (params_idx == 2):
title(r'$\gamma = .5\cdot \sqrt{1+ \Omega}$', fontsize = 32)
if(regime_idx == 3):
xtickNames = setp(ax, xticklabels = method_labels)
setp(xtickNames, rotation=30, fontsize=xlabel_font_size)
ylim((ylower[params_idx], yupper[params_idx]))
ymin, ymax = ylim()
activeticks = linspace(ymin, ymax, 7)[1:-1]
yticks(activeticks, ['%.2f'%tick for tick in activeticks])
if 0 == params_idx:
text(-.25, 0.5, r'$\Omega = %s$' %regime[5:],
horizontalalignment='center',
verticalalignment='center',
transform = ax.transAxes,
rotation='vertical',
size = xlabel_font_size)
tag_loc = 3;
# if (median(param_ests[0]) > param_true):
# tag_loc = 3;
xlim((.0,5.5 ))
t = add_inner_title(ax, chr(ord('A') + subplot_idx), loc=tag_loc,
size=dict(size=ABCD_LABEL_SIZE))
t.patch.set_ec("none")
t.patch.set_alpha(0.5)
file_name = os.path.join(FIGS_DIR, fig_id + 'thetas_est_rel_errors' +'.pdf')
get_current_fig_manager().window.showMaximized()
print 'saving to ', file_name
savefig(file_name)
def postProcessJointBoxPlot(table_file_names, fig_id = '',
Methods = ['Initializer', 'Nelder-Mead', 'Fortet'],
regimeNames = ['superT', 'subT', 'crit', 'superSin']):
analyzers = [];
for table_file_name in table_file_names:
analyzer = DataAnalyzer(table_file_name)
analyzers.append(analyzer)
analyzer = analyzers[0]
if None==regimeNames:
regimeNames = analyzer.getAllRegimes()
for regime in regimeNames:
abg_true = analyzer.getTrueParamValues(regime)
print regime, abg_true
mpl.rcParams['figure.figsize'] = 17, 6
method_labels = ['Initializer', 'Numerical FP', 'Fortet']
xlabel_font_size = 24
sample_size_texts = ['N=100', 'N=1000']
for regime in regimeNames:
yupper = 3*[-inf];
ylower = 3*[inf]
for params_idx in xrange(3):
for mthd in Methods:
for analyzer in analyzers:
ests = array(analyzer.getAllEstimates(regime, mthd))[:,params_idx]
ylower[params_idx] = amin([ylower[params_idx], amin(ests)])
yupper[params_idx] = amax([yupper[params_idx], amax(ests)])
abg_true = analyzer.getTrueParamValues(regime)
print regime, abg_true
figure()
subplots_adjust(left=0.1, right=0.975, top=0.9, bottom=0.25,
wspace = 0.25, hspace = 0.01)
lgs = GridSpec(2, 3)
for sample_size_idx, analyzer in enumerate(analyzers):
abgs = {}
for mthd in Methods:
abgs[mthd] = array(analyzer.getAllEstimates(regime, mthd))
# method_xpos = dict.fromkeys(Methods, [1,2,3])
param_tags = [r' \alpha ', r' \beta ', r' \gamma ']
for params_idx, param_tag in enumerate(param_tags):
print params_idx, param_tag
subplot_idx = sample_size_idx*3 + params_idx;
ax = subplot(lgs[subplot_idx]); hold(True)
# for mthd in Methods:
param_ests = [ abgs[mthd][:,params_idx] for mthd in Methods]
param_true = abg_true[params_idx]
boxplot(param_ests, positions = [1,3,5])
hlines(param_true, 0,5, linestyles='dashed')
ylabel(r'$\hat{%s}$'%param_tag, fontsize = xlabel_font_size)
if (sample_size_idx ==0):
title(r'$%s = %.2f$'%(param_tag, param_true), fontsize = 32)
if(sample_size_idx ==1):
xtickNames = setp(ax, xticklabels = method_labels)
setp(xtickNames, rotation=30, fontsize=xlabel_font_size)
ylim((ylower[params_idx], yupper[params_idx]))
ymin, ymax = ylim()
activeticks = linspace(ymin, ymax, 7)[1:-1]
yticks(activeticks, ['%.2f'%tick for tick in activeticks])
if 0 == params_idx:
text(-.25, 0.5, sample_size_texts[sample_size_idx],
horizontalalignment='center',
verticalalignment='center',
transform = ax.transAxes,
rotation='vertical',
size = xlabel_font_size)
tag_loc = 3;
# if (median(param_ests[0]) > param_true):
# tag_loc = 3;
xlim((.0, 5.5))
t = add_inner_title(ax, chr(ord('A') + subplot_idx), loc=tag_loc,
size=dict(size=ABCD_LABEL_SIZE))
t.patch.set_ec("none")
t.patch.set_alpha(0.5)
file_name = os.path.join(FIGS_DIR, fig_id + regime+'_est_rel_errors_joint' +'.pdf')
get_current_fig_manager().window.showMaximized()
print 'saving to ', file_name
savefig(file_name)
def flagWarnings(table_names):
for table_name in table_names:
analyzer = DataAnalyzer(table_name);
print table_name
print analyzer.getAllWarnings()
def alphaVsGamma(table_file_name, fig_id='',
Methods=['Initializer', 'FP', 'Fortet']):
analyzer = DataAnalyzer(table_file_name)
mpl.rcParams['figure.figsize'] = 17, 5*4
mpl.rcParams['figure.subplot.left'] = .1
mpl.rcParams['figure.subplot.right'] = .975
mpl.rcParams['figure.subplot.bottom'] = .025
mpl.rcParams['figure.subplot.top'] = .95
mpl.rcParams['figure.subplot.wspace'] = .25
mpl.rcParams['figure.subplot.hspace'] = .375
fig = figure();
plot_scale = 1.05;
regimeNames = analyzer.getAllRegimes()
regime_tags = {'superT':'Supra-Threshold', 'crit':'Critical',
'subT':'Sub-Threshold', 'superSin':'Super-Sinusoidal',
'theta1':r'$\Omega=1$',
'theta5':r'$\Omega=5$',
'theta10':r'$\Omega=10$',
'theta20':r'$\Omega=20$'}
xlabel_font_size = 32
for regime_idx,regime in enumerate(regimeNames):
abg_true = analyzer.getTrueParamValues(regime)
print regime, abg_true
for method_idx, method in enumerate(Methods):
abgs = array( analyzer.getAllEstimates(regime, method) )
axu_lim = amax( r_[ abgs[:, 0], abg_true[0]])
axl_lim = amin( r_[ abgs[:, 0], abg_true[0]])
plot_buffer = 5*(plot_scale -1.)*(axu_lim - axl_lim);
axu_lim += plot_buffer
axl_lim -= plot_buffer
ayu_lim = amax( r_[ abgs[:, 2], abg_true[2]])
ayl_lim = amin( r_[ abgs[:, 2], abg_true[2]])
plot_buffer = (plot_scale -1.)*(ayu_lim - ayl_lim);
ayu_lim += plot_buffer
ayl_lim -= plot_buffer
ax = subplot(len(regimeNames),3,
regime_idx*len(Methods) + method_idx+1)
plot(abgs[:, 0], abgs[:, 2],
linestyle='None',
color= 'k', marker= 'o', markersize = 7 )
if 0 == method_idx:
text(-.3, 0.5, regime_tags[regime],
horizontalalignment='center',
verticalalignment='center',
transform = ax.transAxes,
rotation='vertical',
size = xlabel_font_size)
# axu_lim, axl_lim = xlim()
hlines(abg_true[2], axl_lim, axu_lim,
linestyles = 'dashed' )
# axu_lim, axl_lim = ylim()
vlines(abg_true[0], ayl_lim, ayu_lim,
linestyles = 'dashed' )
xlim((axl_lim, axu_lim));
ylim((ayl_lim, ayu_lim));
xlabel(r'$\alpha$',fontsize = xlabel_font_size);
ylabel(r'$\gamma$',fontsize = xlabel_font_size);
title(method , fontsize = 32) #+ ' : $\\Omega= %.1f$' %sqrt((2.0*abg_true[2])**2 -1.)
t = add_inner_title(ax, chr(ord('A') +
regime_idx*len(Methods) + method_idx),
loc=2, size=dict(size=ABCD_LABEL_SIZE))
t.patch.set_ec("none")
t.patch.set_alpha(0.5)
if '' != fig_id:
file_name = os.path.join(FIGS_DIR, fig_id + '_alphagamma_compare_joint.pdf')
print 'saving to ', file_name
savefig(file_name)
def crossAnalyzeJoint(table_file_name1,
table_file_name2,
fig_id = '', Method1 = 'Fortet', Method2 = 'FP',
regimeNames = None):
analyzer = DataAnalyzer(table_file_name1)
analyzer2 = DataAnalyzer(table_file_name2)
mpl.rcParams['figure.figsize'] = 17, 5*4
mpl.rcParams['figure.subplot.left'] = .05
mpl.rcParams['figure.subplot.right'] = .975
mpl.rcParams['figure.subplot.bottom'] = .025
mpl.rcParams['figure.subplot.top'] = .95
mpl.rcParams['figure.subplot.wspace'] = .1
mpl.rcParams['figure.subplot.hspace'] = .375
fig = figure();
if None==regimeNames:
regimeNames = analyzer.getAllRegimes()
for regime_idx,regime in enumerate(regimeNames):
abg_true = analyzer.getTrueParamValues(regime)
print regime, abg_true
abgs = {};
abgs[Method1] = array( analyzer.getAllEstimates(regime, Method1) )
abgs[Method2] = array( analyzer.getAllEstimates(regime, Method2) )
abgs2 = {}
abgs2[Method1] = array( analyzer2.getAllEstimates(regime, Method1) )
abgs2[Method2] = array( analyzer2.getAllEstimates(regime, Method2) )
plot_scale = 1.05;
for param_idx, pname in zip(arange(3),
[r'\alpha', r'\beta', r'\gamma']):
axu_lim = amax( r_[ abgs[Method1][:, param_idx], abgs[Method2][:, param_idx]])
axl_lim = amin( r_[ abgs[Method1][:, param_idx], abgs[Method2][:, param_idx]])
plot_buffer = (plot_scale -1.)*(axu_lim - axl_lim);
axu_lim += plot_buffer
axl_lim -= plot_buffer
ax = subplot(len(regimeNames),3, regime_idx*3 + param_idx+1, aspect='equal')
plot(abgs[Method1][:, param_idx], abgs[Method2][:, param_idx],
linestyle='None',
color= 'k', marker= 'x', markersize = 7, label='$N=100$' )
plot(abgs2[Method1][:, param_idx], abgs2[Method2][:, param_idx],
linestyle='None',
color= 'k', marker= 'd', markersize = 7, label='$N=1000$' )
if param_idx == 0:
legend(loc='lower left')
# axu_lim, axl_lim = xlim()
hlines(abg_true[param_idx], axl_lim, axu_lim,
linestyles = 'dashed' )
# axu_lim, axl_lim = ylim()
vlines(abg_true[param_idx], axl_lim, axu_lim,
linestyles = 'dashed' )
ax.xaxis.set_major_locator(MaxNLocator(6))
ax.yaxis.set_major_locator(MaxNLocator(6))
xlim((axl_lim, axu_lim));
ylim((axl_lim, axu_lim));
xlabel(Method1,fontsize = 24); ylabel(Method2,fontsize = 24);
title('$'+ pname + ' = %.2f$'%abg_true[param_idx], fontsize = 32)
t = add_inner_title(ax, chr(ord('A') +
regime_idx*3 + param_idx),
loc=2, size=dict(size=20))
t.patch.set_ec("none")
t.patch.set_alpha(0.5)
# markers = ['x', '+', '^']:
# for marker_idx, table_file_name in table_file_names[1:]:
# analyzer = DataAnalyzer(table_file_name)
if '' != fig_id:
file_name = os.path.join(FIGS_DIR, fig_id + '_cross_compare_joint.pdf')
print 'saving to ', file_name
savefig(file_name)
def Harness(sample_id=13, regime_name='superSin', N_spikes = 1000, visualize=False):
from scipy.stats.distributions import norm
N_phi = 20;
phi_norms = linspace(1/(2.*N_phi), 1. - 1/ (2.*N_phi), N_phi)
base_name = 'sinusoidal_spike_train_N=%d_'%N_spikes
regime_label = base_name + regime_name
# T_thresh = 128.;
file_name = regime_label + '_' + str(sample_id)
print file_name
binnedTrain = BinnedSpikeTrain.initFromFile(file_name, phi_norms)
# print 'Warning: pruning bins'
# binnedTrain.pruneBins(None, N_thresh = 100)
bins = binnedTrain.bins;
phis = bins.keys()
N_phi = len(phis)
alpha, beta, gamma, theta = binnedTrain._Train._params.getParams()
def loss_function_simple(abg, visualize, fig_tag = ''):
error = .0;
if visualize:
figure()
for (phi_m, phi_idx) in zip(phis, xrange(N_phi)):
Is = bins[phi_m]['Is']
uniqueIs = bins[phi_m]['unique_Is']
a,b,g = abg[0], abg[1], abg[2]
movingThreshold = getMovingThreshold(a,g, phi_m,binnedTrain.theta)
LHS_numerator = movingThreshold(uniqueIs[1:]) *sqrt(2.)
LHS_denominator = b * sqrt(1 - exp(-2*uniqueIs[1:]))
LHS = 1 - norm.cdf(LHS_numerator / LHS_denominator)
RHS = zeros_like(LHS)
N = len(Is)
for rhs_idx in xrange(1,len(uniqueIs)):
t = uniqueIs[rhs_idx]
lIs = Is[Is<t]
taus = t - lIs;
numerator = (movingThreshold(t) - movingThreshold(lIs)* exp(-taus)) * sqrt(2.)
denominator = b * sqrt(1. - exp(-2*taus))
RHS[rhs_idx-1] = sum(1. - norm.cdf(numerator/denominator)) / N
weight = len(Is)
lerror = dot((LHS - RHS)**2 , diff(uniqueIs)) * weight;
error += lerror
if visualize:
subplot(ceil(len(phis)/2),2, phi_idx+1);hold(True)
ts = uniqueIs[1:];
plot(ts, LHS, 'b');
plot(ts, RHS, 'rx');
# annotate('$\phi$ = %.2g'%(phi_m), ((min(ts), max(LHS)/2.)), )
annotate('lerror = %.3g'%lerror,((min(ts), max(LHS)/2.)), )
if visualize:
subplot(ceil(len(phis)/2),2, 1);
title(fig_tag)
return error
def loss_function_nonvectorized(abg, visualize=False):
error = .0;
if visualize:
figure();
for (phi_m, phi_idx) in zip(phis, xrange(N_phi)):
# for phi_m in phis:
Is = bins[phi_m]['Is']
N = len(Is);
Tf = amax(Is);
a,b,g = abg[0], abg[1], abg[2]
movingThreshold = getMovingThreshold(a,g, phi_m)
def LHS(ts):
LHS_numerator = movingThreshold(ts) *sqrt(2.)
LHS_denominator = b * sqrt(1 - exp(-2*ts))
return 1 - norm.cdf(LHS_numerator / LHS_denominator);
# def RHS(t):
# lIs = Is[Is<t];
# taus = t - lIs;
# numerator = (movingThreshold(t) - movingThreshold(lIs)* exp(-taus)) * sqrt(2.)
# denominator = b * sqrt(1. - exp(-2*taus))
# return sum(1. - norm.cdf(numerator/denominator)) / N
def RHS(ts):
if False == iterable(ts):
ts = [ts]
rhs = empty_like(ts)
for t, t_indx in zip(ts, xrange(size(ts))):
lIs = Is[Is<t];
taus = t - lIs;
numerator = (movingThreshold(t) - movingThreshold(lIs)* exp(-taus)) * sqrt(2.)
denominator = b * sqrt(1. - exp(-2*taus))
rhs[t_indx] = sum(1. - norm.cdf(numerator/denominator)) / N
return rhs
integrand = lambda t: (LHS(t) - RHS(t)) **2
from scipy.integrate import quad, quadrature, fixed_quad
# quadrature, quad_error = quad(integrand, a= 1e-8, b=Tf+1e-8, limit = 50)
quadrature, quad_error = quadrature(integrand, a= 1e-8, b=Tf+1e-8,
tol=5e-03, rtol=1.49e-04,
maxiter = 64,
vec_func = True)
# val , err_msg = fixed_quad( integrand, a= 1e-8, b=Tf+1e-8,
# n = 12)
# print 'quadrature = ',quadrature
# print 'val = ',val
# print 'difference = ', quadrature - val
weight = len(Is)
#VISUALIZE FOR NOW:
if visualize:
subplot(ceil(len(phis)/2),2, phi_idx+1);hold(True)
ts = linspace(1e-8, Tf+1e-8, 100) ;
lhs = empty_like(ts); rhs = empty_like(ts);
for t, t_indx in zip(ts,
xrange(len(ts))):
lhs[t_indx] = LHS(t);
rhs[t_indx] = RHS(t);
plot(ts, lhs, 'b');
plot(ts, rhs, 'rx');
error += quadrature* weight;
return error
def loss_function_quadGaussian(abg, visualize=False, fig_tag = ''):
error = .0;
if visualize:
figure();
for (phi_m, phi_idx) in zip(phis, xrange(N_phi)):
# for phi_m in phis:
Is = bins[phi_m]['Is']
unique_Is = bins[phi_m]['unique_Is']
N_Is = len(Is);
Tf = amax(Is);
a,b,g = abg[0], abg[1], abg[2]
movingThreshold = getMovingThreshold(a,g, phi_m)
def LHS(ts):
LHS_numerator = movingThreshold(ts) *sqrt(2.)
LHS_denominator = b * sqrt(1 - exp(-2*ts))
return 1 - norm.cdf(LHS_numerator / LHS_denominator);
def RHS(ts):
if False == iterable(ts):
ts = array([ts])
# rhs = empty_like(ts)
# Is.reshape((len(Is),1))
lIs = tile(Is, len(ts) ).reshape((len(ts), len(Is))).transpose()
lts = tile(ts, (len(Is),1 ) )
mask = lIs < lts
taus = (lts - lIs); #*mask
#NOTE BELOW WE use abs(taus) since for non-positive taus we will mask away anyway:
numerator = (movingThreshold(lts) - movingThreshold(lIs)* exp(-abs(taus))) * sqrt(2.)
denominator = b * sqrt(1. - exp(-2*abs(taus)))
rhs = sum( (1. - norm.cdf(numerator/denominator))*mask, axis=0) / N_Is
return rhs
integrand = lambda t: (LHS(t) - RHS(t)) **2
from scipy.integrate import quad, quadrature, fixed_quad
# valcheck, quad_error = quad(integrand, a= 1e-8, b=Tf+1e-8, limit = 64)
val, quad_error = quadrature(integrand, a= 1e-8, b=Tf+1e-8,
tol=5e-03, rtol=1.49e-04,
maxiter = 64,
vec_func = True)
weight = len(Is)
lerror = val* weight;
error += lerror
if visualize:
subplot(ceil(len(phis)/2),2, phi_idx+1); hold(True)
# ts = linspace(1e-8, Tf+1e-8, 100) ;
ts = unique_Is[1:];
lhs = LHS(ts);
rhs = RHS(ts);
plot(ts, lhs, 'b');
plot(ts, rhs, 'rx');
annotate('lerror = %.3g'%lerror,((min(ts), max(lhs)/2.)), )
if visualize:
subplot(ceil(len(phis)/2),2,1);
title(fig_tag)
return error
def loss_function_L1(abg, visualize=False, fig_tag = ''):
error = .0;
if visualize:
figure();
for (phi_m, phi_idx) in zip(phis, xrange(N_phi)):
# for phi_m in phis:
Is = bins[phi_m]['Is']
unique_Is = bins[phi_m]['unique_Is']
N_Is = len(Is);
Tf = amax(Is);
a,b,g = abg[0], abg[1], abg[2]
movingThreshold = getMovingThreshold(a,g, phi_m, theta)
def LHS(ts):
LHS_numerator = movingThreshold(ts) *sqrt(2.)
LHS_denominator = b * sqrt(1 - exp(-2*ts))
return 1 - norm.cdf(LHS_numerator / LHS_denominator);
def RHS(ts):
if False == iterable(ts):
ts = array([ts])
# rhs = empty_like(ts)
# Is.reshape((len(Is),1))
lIs = tile(Is, len(ts) ).reshape((len(ts), len(Is))).transpose()
lts = tile(ts, (len(Is),1 ) )
mask = lIs < lts
taus = (lts - lIs); #*mask
#NOTE BELOW WE use abs(taus) since for non-positive taus we will mask away anyway:
numerator = (movingThreshold(lts) - movingThreshold(lIs)* exp(-abs(taus))) * sqrt(2.)
denominator = b * sqrt(1. - exp(-2*abs(taus)))
rhs = sum( (1. - norm.cdf(numerator/denominator))*mask, axis=0) / N_Is
return rhs
integrand = lambda t: abs(LHS(t) - RHS(t))
from scipy.integrate import quad, quadrature, fixed_quad
print unique_Is
val, quad_error = quad(integrand, a= 1e-8, b=Tf+1., limit = 64,
points = sort(unique_Is) )
# val, quad_error = quadrature(integrand, a= 1e-8, b=Tf+1e-8,
# tol=5e-03, rtol=1.49e-04,
# maxiter = 64,
# vec_func = True)
# val , err_msg = fixed_quad( integrand, a= 1e-8, b=Tf+1e-8,
weight = len(Is)
lerror = val* weight;
error += lerror
if visualize:
subplot(ceil(len(phis)/2),2, phi_idx+1); hold(True)
# ts = linspace(1e-8, Tf+1e-8, 100) ;
ts = unique_Is[1:];
lhs = LHS(ts);
rhs = RHS(ts);
plot(ts, lhs, 'b');
plot(ts, rhs, 'rx');
annotate('lerror = %.3g'%lerror,((min(ts), max(lhs)/2.)), )
if visualize:
subplot(ceil(len(phis)/2),2,1); title(fig_tag)
return error
def loss_function_manualquad(abg, visualize=False, fig_tag = ''):
error = .0;
if visualize:
figure();
for (phi_m, phi_idx) in zip(phis, xrange(N_phi)):
Is = bins[phi_m]['Is']
unique_Is = bins[phi_m]['unique_Is']
N_Is = len(Is);
Tf = amax(Is);
a,b,g = abg[0], abg[1], abg[2]
movingThreshold = getMovingThreshold(a,g, phi_m, binnedTrain.theta)
def LHS(ts):
LHS_numerator = movingThreshold(ts) *sqrt(2.)
LHS_denominator = b * sqrt(1 - exp(-2*ts))
return 1 - norm.cdf(LHS_numerator / LHS_denominator);
def RHS(ts):
if False == iterable(ts):
ts = array([ts])
lIs = tile(Is, len(ts) ).reshape((len(ts), len(Is))).transpose()
lts = tile(ts, (len(Is),1 ) )
mask = lIs < lts
taus = (lts - lIs); #*mask
#NOTE BELOW WE use abs(taus) since for non-positive taus we will mask away anyway:
numerator = (movingThreshold(lts) - movingThreshold(lIs)* exp(-abs(taus))) * sqrt(2.)
denominator = b * sqrt(1. - exp(-2*abs(taus)))
rhs = sum( (1. - norm.cdf(numerator/denominator))*mask, axis=0) / N_Is
return rhs
integrand = lambda t: (LHS(t) - RHS(t)) **2
dt = 1e-2;
ts_manual = arange(1e-8, Tf+1e-2, dt )
integrand = LHS(ts_manual) - RHS(ts_manual)
val = dot(integrand, integrand)*dt;
weight = len(Is)
lerror = val* weight;
error += lerror
if visualize:
subplot(ceil(len(phis)/2),2, phi_idx+1); hold(True)
# ts = linspace(1e-8, Tf+1e-8, 100) ;
ts = unique_Is[1:];
lhs = LHS(ts);
rhs = RHS(ts);
plot(ts, lhs, 'b');
plot(ts, rhs, 'rx');
annotate('lerror = %.3g'%lerror,((min(ts), max(lhs)/2.)), )
if visualize:
subplot(ceil(len(phis)/2),2,1);
title(fig_tag)
return error
def loss_function_supnormalized(abg, visualize=False, fig_tag = ''):
error = .0;
if visualize:
figure();
for (phi_m, phi_idx) in zip(phis, xrange(N_phi)):
Is = bins[phi_m]['Is']
unique_Is = bins[phi_m]['unique_Is']
N_Is = len(Is);
Tf = amax(Is);
a,b,g = abg[0], abg[1], abg[2]
movingThreshold = getMovingThreshold(a,g, phi_m, binnedTrain.theta)
def LHS(ts):
LHS_numerator = movingThreshold(ts) *sqrt(2.)
LHS_denominator = b * sqrt(1 - exp(-2*ts))
return 1 - norm.cdf(LHS_numerator / LHS_denominator);
def RHS(ts):
if False == iterable(ts):
ts = array([ts])
lIs = tile(Is, len(ts) ).reshape((len(ts), len(Is))).transpose()
lts = tile(ts, (len(Is),1 ) )
mask = lIs < lts
taus = (lts - lIs); #*mask
#NOTE BELOW WE use abs(taus) since for non-positive taus we will mask away anyway:
numerator = (movingThreshold(lts) - movingThreshold(lIs)* exp(-abs(taus))) * sqrt(2.)
denominator = b * sqrt(1. - exp(-2*abs(taus)))
rhs = sum( (1. - norm.cdf(numerator/denominator))*mask, axis=0) / N_Is
return rhs
dt = 1e-3;
ts_manual = arange(1e-8, Tf+1e-2, dt)
lhs = LHS(ts_manual)
difference = abs(lhs - RHS(ts_manual))/amax(lhs)
val = amax(difference);
weight = len(Is)
lerror = val* weight;
error += lerror
if visualize:
subplot(ceil(len(phis)/2),2, phi_idx+1); hold(True)
# ts = linspace(1e-8, Tf+1e-8, 100) ;
ts = unique_Is[1:];
lhs = LHS(ts);
rhs = RHS(ts);
plot(ts, lhs, 'b');
plot(ts, rhs, 'rx');
annotate('lerror = %.3g'%lerror,((min(ts), max(lhs)/2.)), )
if visualize:
subplot(ceil(len(phis)/2),2,1);
title(fig_tag)
return error
def loss_function_sup(abg, visualize=False, fig_tag = ''):
error = .0;
if visualize:
figure();
for (phi_m, phi_idx) in zip(phis, xrange(N_phi)):
Is = bins[phi_m]['Is']
unique_Is = bins[phi_m]['unique_Is']
N_Is = len(Is);
Tf = amax(Is);
a,b,g = abg[0], abg[1], abg[2]
movingThreshold = getMovingThreshold(a,g, phi_m, binnedTrain.theta)
def LHS(ts):
LHS_numerator = movingThreshold(ts) *sqrt(2.)
LHS_denominator = b * sqrt(1 - exp(-2*ts))
return 1 - norm.cdf(LHS_numerator / LHS_denominator);
def RHS(ts):
if False == iterable(ts):
ts = array([ts])
lIs = tile(Is, len(ts) ).reshape((len(ts), len(Is))).transpose()
lts = tile(ts, (len(Is),1 ) )
mask = lIs < lts
taus = (lts - lIs); #*mask
#NOTE BELOW WE use abs(taus) since for non-positive taus we will mask away anyway:
numerator = (movingThreshold(lts) - movingThreshold(lIs)* exp(-abs(taus))) * sqrt(2.)
denominator = b * sqrt(1. - exp(-2*abs(taus)))
rhs = sum( (1. - norm.cdf(numerator/denominator))*mask, axis=0) / N_Is
return rhs
dt = 1e-3;
ts_manual = arange(1e-8, Tf+1e-2, dt)
lhs = LHS(ts_manual)
difference = abs(lhs - RHS(ts_manual))
val = amax(difference);
weight = len(Is)
lerror = val* weight;
error += lerror
if visualize:
subplot(ceil(len(phis)/2),2, phi_idx+1); hold(True)
# ts = linspace(1e-8, Tf+1e-8, 100) ;
ts = unique_Is[1:];
lhs = LHS(ts);
rhs = RHS(ts);
plot(ts, lhs, 'b');
plot(ts, rhs, 'rx');
annotate('lerror = %.3g'%lerror,((min(ts), max(lhs)/2.)), )
if visualize:
subplot(ceil(len(phis)/2),2,1);
title(fig_tag)
return error
#EXPERIMENT:
# Analyzer = DataAnalyzer()
def outlinept():
pass
analyzer = DataAnalyzer('FvsWF_4x16');
abg_true = analyzer.getTrueParamValues(regime_name)
loss_function_L1(abg_true, visualize=True)
return
quad_estimated = analyzer.getEstimates(sample_id, regime_name, 'QuadFortet')[0]
simple_estimated = analyzer.getEstimates(sample_id, regime_name, 'Fortet')[0]
for abg, tag, L in zip(3*[abg_true, quad_estimated],
['sup_true_params' , 'sup_estimated_params',
'supnormailzed_true_params', 'supnormailzed_estimated_params',
'manualquad_true_params' , 'manualquad_estimated_params'],
2*[loss_function_sup]+
2*[loss_function_supnormalized] +
2*[loss_function_manualquad]):
start = time.clock()
loss = L(abg,visualize, fig_tag = regime + '_' + tag);
end = time.clock()
print tag, ':%.2f,%.2f,%.2f:' %(abg[0],abg[1],abg[2]), 'error = %.4f'%loss , ' | compute time = ', end - start
