def marginalized_analysis(): sys.stdout = tee.tee(STDOUT, open(OUTDIR+'eg-appendix1c.out', 'w')) x, y = make_data() prior = gv.gvar(91 * ['0(1)']) # prior for the fit ymod = y - (f(x, prior) - f(x, prior[:1])) priormod = prior[:1] fit = lsqfit.nonlinear_fit(data=(x, ymod), prior=priormod, fcn=f) print fit.format(maxline=True) sys.stdout = STDOUT print lsqfit.wavg(list(ymod) + list(priormod)) make_plot(x, ymod, fit, 'ymod(x)') inputs = dict(prior=prior, y0=y[0], y1=y[1], y2=y[2], y3=y[3], y4=y[4]) outputs = dict(p0=fit.p[0]) print gv.fmt_errorbudget(inputs=inputs, outputs=outputs) return fit
def marginalized_analysis():
sys.stdout = tee.tee(STDOUT, open(OUTDIR + 'eg-appendix1c.out', 'w'))
x, y = make_data()
prior = gv.gvar(91 * ['0(1)']) # prior for the fit
ymod = y - (f(x, prior) - f(x, prior[:1]))
priormod = prior[:1]
fit = lsqfit.nonlinear_fit(data=(x, ymod), prior=priormod, fcn=f)
print fit.format(maxline=True)
sys.stdout = STDOUT
print lsqfit.wavg(list(ymod) + list(priormod))
make_plot(x, ymod, fit, 'ymod(x)', name='eg-appendix1c')
inputs = dict(prior=prior, y0=y[0], y1=y[1], y2=y[2], y3=y[3], y4=y[4])
outputs = dict(p0=fit.p[0])
print gv.fmt_errorbudget(inputs=inputs, outputs=outputs)
return fit
def good_analysis(): sys.stdout = tee.tee(STDOUT, open(OUTDIR+'eg-appendix1b.out', 'w')) x, y = make_data() prior = gv.gvar(N * ['0(1)']) # prior for the fit fit = lsqfit.nonlinear_fit(data=(x, y), prior=prior, fcn=f) print fit.format(maxline=True) make_plot(x, y, fit) inputs = gv.BufferDict(prior=prior) for xi, yi in zip(x, y): inputs['y(%.2f)' % xi] = yi sys.stdout = tee.tee(STDOUT, open(OUTDIR+'eg-appendix1g.out', 'w')) inputs = dict(prior=prior, y=y) outputs = dict(p0=fit.p[0]) print gv.fmt_errorbudget(inputs=inputs, outputs=outputs) return fit
def main():
x, y = make_data() # collect fit data
p0 = None # make larger fits go faster (opt.)
for nexp in range(1, 7):
print('************************************* nexp =', nexp)
prior = make_prior(nexp)
fit = lsqfit.nonlinear_fit(data=(x, y), fcn=fcn, prior=prior, p0=p0)
print(fit) # print the fit results
if nexp > 2:
E = fit.p['E'] # best-fit parameters
a = fit.p['a']
print('E1/E0 =', E[1] / E[0], ' E2/E0 =', E[2] / E[0])
print('a1/a0 =', a[1] / a[0], ' a2/a0 =', a[2] / a[0])
if fit.chi2 / fit.dof < 1.:
p0 = fit.pmean # starting point for next fit (opt.)
print()
# error budget analysis
# outputs = {
# 'E1/E0':E[1]/E[0], 'E2/E0':E[2]/E[0],
# 'a1/a0':a[1]/a[0], 'a2/a0':a[2]/a[0]
# }
# inputs = {'E':fit.prior['E'], 'a':fit.prior['a'], 'y':y}
outputs = gv.BufferDict()
outputs['E2/E0'] = E[2] / E[0]
outputs['E1/E0'] = E[1] / E[0]
outputs['a2/a0'] = a[2] / a[0]
outputs['a1/a0'] = a[1] / a[0]
inputs = gv.BufferDict()
inputs['a'] = fit.prior['a']
inputs['y'] = y
inputs['E'] = fit.prior['E']
print('================= Error Budget Analysis')
print(gv.fmt_values(outputs))
print(gv.fmt_errorbudget(outputs, inputs))
def good_analysis(plot=MAKE_PLOTS):
sys.stdout = tee.tee(STDOUT, open(OUTDIR + 'eg-appendix1b.out', 'w'))
x, y = make_data()
prior = gv.gvar(N * ['0(1)']) # prior for the fit
fit = lsqfit.nonlinear_fit(data=(x, y), prior=prior, fcn=f)
print fit.format(maxline=True)
if plot:
make_plot(x, y, fit, name='eg-appendix1b')
inputs = gv.BufferDict(prior=prior)
for xi, yi in zip(x, y):
inputs['y(%.2f)' % xi] = yi
sys.stdout = tee.tee(STDOUT, open(OUTDIR + 'eg-appendix1g.out', 'w'))
inputs = dict(prior=prior, y=y)
outputs = dict(p0=fit.p[0])
print gv.fmt_errorbudget(inputs=inputs, outputs=outputs)
return fit
def write_results(fit, basis, prior, data, N):
l.write('Parameters used: ' + '\n')
l.write('file: '+ file + '\n')
l.write('t0: '+ str(t0) + '\n')
l.write('T: '+ str(T) + '\n')
l.write('tmin: '+ str(tmin) + '\n')
l.write('tmax: '+ str(tmax) + '\n\n')
l.write('offtmin: '+ str(offtmin) + '\n')
l.write('offtmax: '+ str(offtmax) + '\n\n')
l.write(30 * '=' + '\n' + 'nterm = ' + str(N[-1]) + '\n')
l.write(fit.format(pstyle=None if N < 7 else 'm'))
l.write(30 * '=' + ' Results\n')
l.write(basis.tabulate(fit.p, keyfmt=tag+'{s1}'))
if OSC:
l.write(basis.tabulate(fit.p, keyfmt=otag+'{s1}'))
E = np.cumsum(fit.p[tag+'dE'])
outputs = collections.OrderedDict()
outputs['a*E(2s-1s)'] = E[1] - E[0]
outputs['a*E(3s-1s)'] = E[2] - E[0]
outputs['E(3s-1s)/E(2s-1s)'] = (E[2] - E[0]) / (E[1] - E[0])
inputs = collections.OrderedDict()
inputs['prior'] = prior
inputs['data'] = data
inputs['svdcut'] = fit.svdcorrection
l.write(gv.fmt_values(outputs))
l.write(gv.fmt_errorbudget(outputs, inputs, colwidth=18))
l.write('Prior:\n')
for k in [tag+SRC[0], tag+SRC[1]]:
l.write('{:8}{}\n'.format(k, list(prior[k])))
def print_results(fit, prior, data):
""" Print results of fit. """
outputs = collections.OrderedDict()
outputs['mDs'] = fit.p['dE'][0]
outputs['Vnn'] = fit.p['Vnn'][0]
inputs = collections.OrderedDict()
inputs['statistics'] = data # statistical errors in data
inputs.update(prior) # errors from priors
inputs['svd'] = fit.svdcorrection # errors from svd cut (if present)
print('\n' + gv.fmt_values(outputs))
print(gv.fmt_errorbudget(outputs, inputs))
def main():
param, data = collect_data('spline.p')
F, prior = make_fcn_prior(param)
fit = lsqfit.nonlinear_fit(data=data, prior=prior, fcn=F)
print(fit)
# create f(m)
f = gv.cspline.CSpline(fit.p['mknot'], fit.p['fknot'])
# create error budget
outputs = {'f(1)':f(1), 'f(5)':f(5), 'f(9)':f(9)}
inputs = {'data':data}
inputs.update(prior)
print(gv.fmt_values(outputs))
print(gv.fmt_errorbudget(outputs=outputs, inputs=inputs))
make_plot(param, data, fit)
def print_results(fit, basis, prior, data):
print(30 * '=', 'Results\n')
print(basis.tabulate(fit.p, keyfmt='etab.{s1}'))
print(basis.tabulate(fit.p, keyfmt='etab.{s1}', eig_srcs=True))
E = np.cumsum(fit.p['etab.dE'])
outputs = collections.OrderedDict()
outputs['a*E(2s-1s)'] = E[1] - E[0]
outputs['a*E(3s-1s)'] = E[2] - E[0]
outputs['E(3s-1s)/E(2s-1s)'] = (E[2] - E[0]) / (E[1] - E[0])
inputs = collections.OrderedDict()
inputs['prior'] = prior
inputs['data'] = data
inputs['svdcut'] = fit.correction
print(gv.fmt_values(outputs))
print(gv.fmt_errorbudget(outputs, inputs, colwidth=18))
def print_results(fit, basis, prior, data):
print(30 * '=', 'Results\n')
print(basis.tabulate(fit.p, keyfmt='etab.{s1}'))
print(basis.tabulate(fit.p, keyfmt='etab.{s1}', eig_srcs=True))
E = np.cumsum(fit.p['etab.dE'])
outputs = collections.OrderedDict()
outputs['a*E(2s-1s)'] = E[1] - E[0]
outputs['a*E(3s-1s)'] = E[2] - E[0]
outputs['E(3s-1s)/E(2s-1s)'] = (E[2] - E[0]) / (E[1] - E[0])
inputs = collections.OrderedDict()
inputs['prior'] = prior
inputs['data'] = data
inputs['svdcut'] = fit.svdcorrection
print(gv.fmt_values(outputs))
print(gv.fmt_errorbudget(outputs, inputs, colwidth=18))
def print_results(fit, prior, data):
""" Print results of fit. """
outputs = collections.OrderedDict()
outputs['mDs'] = fit.p['dE'][0]
outputs['Vnn'] = fit.p['Vnn'][0]
inputs = collections.OrderedDict()
inputs['statistics'] = data # statistical errors in data
inputs['Ds priors'] = {
k:prior[k] for k in ['log(a)', 'log(dE)', 'log(ao)', 'log(dEo)']
}
inputs['V priors'] = {
k:prior[k] for k in ['Vnn', 'Vno', 'Voo']
}
print('\n' + gv.fmt_values(outputs))
print(gv.fmt_errorbudget(outputs, inputs))
def print_results(fit, prior, data):
""" Report best-fit results. """
print('Fit results:')
p = fit.p # best-fit parameters
# etas
E_etas = np.cumsum(p['etas:dE'])
a_etas = p['etas:a']
print(' Eetas:', E_etas[:3])
print(' aetas:', a_etas[:3])
# Ds
E_Ds = np.cumsum(p['Ds:dE'])
a_Ds = p['Ds:a']
print('\n EDs:', E_Ds[:3])
print( ' aDs:', a_Ds[:3])
# Dso -- oscillating piece
E_Dso = np.cumsum(p['Dso:dE'])
a_Dso = p['Dso:a']
print('\n EDso:', E_Dso[:3])
print( ' aDso:', a_Dso[:3])
# V
Vnn = p['Vnn']
Vno = p['Vno']
print('\n etas->V->Ds =', Vnn[0, 0])
print(' etas->V->Dso =', Vno[0, 0])
# error budget
outputs = collections.OrderedDict()
outputs['metas'] = E_etas[0]
outputs['mDs'] = E_Ds[0]
outputs['mDso-mDs'] = E_Dso[0] - E_Ds[0]
outputs['Vnn'] = Vnn[0, 0]
outputs['Vno'] = Vno[0, 0]
inputs = collections.OrderedDict()
inputs['statistics'] = data # statistical errors in data
inputs.update(prior) # all entries in prior
inputs['svd'] = fit.svdcorrection # svd cut (if present)
print('\n' + gv.fmt_values(outputs))
print(gv.fmt_errorbudget(outputs, inputs))
print('\n')
def main():
x, y = make_data()
prior = make_prior(100) # 100 exponential terms in all
p0 = None
for nexp in range(1, 6):
# marginalize the last 100 - nexp terms (in ymod_prior)
fit_prior = gv.BufferDict() # part of prior used in fit
ymod_prior = gv.BufferDict() # part of prior absorbed in ymod
for k in prior:
fit_prior[k] = prior[k][:nexp]
ymod_prior[k] = prior[k][nexp:]
ymod = y - fcn(x, ymod_prior) # remove temrs in ymod_prior
# fit modified data with just nexp terms (in fit_prior)
fit = lsqfit.nonlinear_fit(
data=(x, ymod), prior=fit_prior, fcn=fcn, p0=p0, tol=1e-15, svdcut=1e-12,
)
# print fit information
print('************************************* nexp =',nexp)
print(fit.format(True))
p0 = fit.pmean
# print summary information and error budget
E = fit.p['E'] # best-fit parameters
a = fit.p['a']
# outputs = {
# 'E1/E0':E[1] / E[0], 'E2/E0':E[2] / E[0],
# 'a1/a0':a[1] / a[0], 'a2/a0':a[2] / a[0]
# }
# inputs = {
# 'E prior':prior['E'], 'a prior':prior['a'],
# 'svd cut':fit.correction,
# }
outputs = gv.BufferDict()
outputs['E2/E0'] = E[2] / E[0]
outputs['E1/E0'] = E[1] / E[0]
outputs['a2/a0'] = a[2] / a[0]
outputs['a1/a0'] = a[1] / a[0]
inputs = gv.BufferDict()
inputs['E prior'] = prior['E']
inputs['svd cut'] = fit.correction
inputs['a prior'] = prior['a']
print(gv.fmt_values(outputs))
print(gv.fmt_errorbudget(outputs, inputs))
def main():
l = gv.gvar(0.25, 0.0005) # length of pendulum
theta_max = gv.gvar(np.pi / 6, 0.025) # max angle of swing
y = make_pendulum(theta_max, l) # y(t) = [theta(t), d/dt theta(t)]
T = find_period(y, Tapprox=1.0)
print('period T = {} sec'.format(T))
fmt = 'uncertainty = {:.2f} min/day\n'
print(fmt.format((T.sdev / T.mean) * 60. * 24.))
# error budget for T
inputs = dict(l=l, theta_max=theta_max)
outputs = {'T':T}
print(gv.fmt_errorbudget(outputs=outputs, inputs=inputs))
# check errors in T using a simulation
Tlist = []
for i in range(200):
y = make_pendulum(theta_max(), l())
Tlist.append(find_period(y, Tapprox=1.0))
print('period T = {:.4f} +- {:.4f}'.format(np.mean(Tlist), np.std(Tlist)))
def print_results(fit, prior, data):
""" Report best-fit results. """
print('Fit results:')
p = fit.p # best-fit parameters
# etas
E_etas = np.cumsum(p['etas:dE'])
a_etas = p['etas:a']
print(' Eetas:', E_etas[:3])
print(' aetas:', a_etas[:3])
# Ds
E_Ds = np.cumsum(p['Ds:dE'])
a_Ds = p['Ds:a']
print('\n EDs:', E_Ds[:3])
print(' aDs:', a_Ds[:3])
# Dso -- oscillating piece
E_Dso = np.cumsum(p['Dso:dE'])
a_Dso = p['Dso:a']
print('\n EDso:', E_Dso[:3])
print(' aDso:', a_Dso[:3])
# V
Vnn = p['Vnn']
print('\n etas->V->Ds =', Vnn[0, 0])
# error budget
outputs = collections.OrderedDict()
outputs['metas'] = E_etas[0]
outputs['mDs'] = E_Ds[0]
outputs['Vnn'] = Vnn[0, 0]
inputs = collections.OrderedDict()
inputs['statistics'] = data # statistical errors in data
inputs['svd'] = fit.correction
inputs.update(prior) # all entries in prior
print('\n' + gv.fmt_values(outputs))
print(gv.fmt_errorbudget(outputs, inputs))
print('\n')
def print_results(fit, basis, prior, data):
print(30 * '=', 'Results\n')
print(basis.tabulate(fit.p, keyfmt=tag+'{s1}'))
if OSC:
print(basis.tabulate(fit.p, keyfmt=otag+'{s1}'))
#print(basis.tabulate(fit.p, keyfmt=tag+'{s1}', eig_srcs=True))
E = np.cumsum(fit.p[tag+'dE'])
outputs = collections.OrderedDict()
outputs['a*E(2s-1s)'] = E[1] - E[0]
outputs['a*E(3s-1s)'] = E[2] - E[0]
outputs['E(3s-1s)/E(2s-1s)'] = (E[2] - E[0]) / (E[1] - E[0])
inputs = collections.OrderedDict()
inputs['prior'] = prior
inputs['data'] = data
inputs['svdcut'] = fit.svdcorrection
print(gv.fmt_values(outputs))
print(gv.fmt_errorbudget(outputs, inputs, colwidth=18))
print('Prior:\n')
for k in [tag+SRC[0], tag+SRC[1]]:
print('{:13}{}'.format(k, list(prior[k])))
print()
def main():
l = gv.gvar(0.25, 0.0005) # length of pendulum
theta_max = gv.gvar(np.pi / 6, 0.025) # max angle of swing
y = make_pendulum(theta_max, l) # y(t) = [theta(t), d/dt theta(t)]
T = find_period(y, Tapprox=1.0)
print('period T = {} sec'.format(T))
fmt = 'uncertainty = {:.2f} min/day\n'
print(fmt.format((T.sdev / T.mean) * 60. * 24.))
# error budget for T
inputs = gv.BufferDict() # dict(l=l, theta_max=theta_max)
inputs['l'] = l
inputs['theta_max'] = theta_max
outputs = {'T': T}
print(gv.fmt_errorbudget(outputs=outputs, inputs=inputs))
# check errors in T using a simulation
Tlist = []
for i in range(200):
y = make_pendulum(theta_max(), l())
Tlist.append(find_period(y, Tapprox=1.0))
print('period T = {:.4f} +- {:.4f}'.format(np.mean(Tlist), np.std(Tlist)))
def print_results(fit, basis, prior, data):
print(30 * '=', 'Results\n')
print(basis.tabulate(fit.p, keyfmt='etab.{s1}'))
print(basis.tabulate(fit.p, keyfmt='etab.{s1}', eig_srcs=True))
E = np.cumsum(fit.p['etab.dE'])
outputs = collections.OrderedDict()
outputs['a*E(2s-1s)'] = E[1] - E[0]
outputs['a*E(3s-1s)'] = E[2] - E[0]
outputs['E(3s-1s)/E(2s-1s)'] = (E[2] - E[0]) / (E[1] - E[0])
inputs = collections.OrderedDict()
inputs['prior'] = prior
inputs['data'] = data
inputs['svdcut'] = fit.correction
print(gv.fmt_values(outputs))
print(gv.fmt_errorbudget(outputs, inputs, colwidth=18))
print('Prior:\n')
for k in ['etab.l', 'etab.g', 'etab.d', 'etab.e']:
print('{:13}{}'.format(k, list(prior[k])))
print()
prior_eig = basis.apply(prior, keyfmt='etab.{s1}')
for k in ['etab.0', 'etab.1', 'etab.2', 'etab.3']:
print('{:13}{}'.format(k, list(prior_eig[k])))
def main():
x, y = make_data() # make fit data
# y = gv.gvar(gv.mean(y), 0.75**2 * gv.evalcov(y))
p0 = None # make larger fits go faster (opt.)
sys_stdout = sys.stdout
sys.stdout = tee.tee(sys.stdout, open("eg1.out", "w"))
for nexp in range(1, 7):
prior = make_prior(nexp)
fit = lsqfit.nonlinear_fit(data=(x, y), fcn=fcn, prior=prior, p0=p0)
if fit.chi2 / fit.dof < 1.:
p0 = fit.pmean # starting point for next fit (opt.)
print('************************************* nexp =', nexp)
print(fit.format()) # print the fit results)
E = fit.p['E'] # best-fit parameters
a = fit.p['a']
if nexp > 2:
print('E1/E0 =', E[1] / E[0], ' E2/E0 =', E[2] / E[0])
print('a1/a0 =', a[1] / a[0], ' a2/a0 =', a[2] / a[0])
print
# error budget
outputs = {
'E1/E0': E[1] / E[0],
'E2/E0': E[2] / E[0],
'a1/a0': a[1] / a[0],
'a2/a0': a[2] / a[0]
}
inputs = {'E': fit.prior['E'], 'a': fit.prior['a'], 'y': y}
inputs = collections.OrderedDict()
inputs['a'] = fit.prior['a']
inputs['E'] = fit.prior['E']
inputs['y'] = fit.data[1]
print('================= Error Budget Analysis')
print(gv.fmt_values(outputs))
print(gv.fmt_errorbudget(outputs, inputs))
sys.stdout = sys_stdout
# print(gv.gvar(str(a[1])) / gv.gvar(str(a[0])) )
# print(gv.evalcorr([fit.p['a'][1], fit.p['E'][1]]))
# print(fit.format(True))
# redo fit with 4 parameters since that is enough
prior = make_prior(4)
fit = lsqfit.nonlinear_fit(data=(x, y), fcn=fcn, prior=prior, p0=fit.pmean)
sys.stdout = tee.tee(sys_stdout, open("eg1a.out", "w"))
print('--------------------- original fit')
print(fit.format())
E = fit.p['E'] # best-fit parameters
a = fit.p['a']
print('E1/E0 =', E[1] / E[0], ' E2/E0 =', E[2] / E[0])
print('a1/a0 =', a[1] / a[0], ' a2/a0 =', a[2] / a[0])
print
# extra data 1
print('\n--------------------- new fit to extra information')
def ratio(p):
return p['a'][1] / p['a'][0]
newfit = lsqfit.nonlinear_fit(data=gv.gvar(1, 1e-5),
fcn=ratio,
prior=fit.p)
print((newfit.format()))
E = newfit.p['E']
a = newfit.p['a']
print('E1/E0 =', E[1] / E[0], ' E2/E0 =', E[2] / E[0])
print('a1/a0 =', a[1] / a[0], ' a2/a0 =', a[2] / a[0])
if DO_PLOT:
import matplotlib.pyplot as plt
ratio = y / fit.fcn(x, fit.pmean)
plt.xlim(4, 15)
plt.ylim(0.95, 1.05)
plt.xlabel('x')
plt.ylabel('y / f(x,p)')
plt.yticks([0.96, 0.98, 1.00, 1.02, 1.04],
['0.96', '0.98', '1.00', '1.02', '1.04'])
plt.errorbar(x=x, y=gv.mean(ratio), yerr=gv.sdev(ratio), fmt='ob')
plt.plot([4.0, 21.0], [1.0, 1.0], 'b:')
plt.savefig('eg1.png', bbox_inches='tight')
plt.show()
# alternate method for extra data
sys.stdout = tee.tee(sys_stdout, open("eg1b.out", "w"))
fit.p['a1/a0'] = fit.p['a'][1] / fit.p['a'][0]
new_data = {'a1/a0': gv.gvar(1, 1e-5)}
new_p = lsqfit.wavg([fit.p, new_data])
print('chi2/dof = %.2f\n' % (new_p.chi2 / new_p.dof))
print('E:', new_p['E'][:4])
print('a:', new_p['a'][:4])
print('a1/a0:', new_p['a1/a0'])
if DO_BAYES:
# Bayesian Fit
gv.ranseed([123])
prior = make_prior(4)
fit = lsqfit.nonlinear_fit(data=(x, y),
fcn=f,
prior=prior,
p0=fit.pmean)
sys.stdout = tee.tee(sys_stdout, open("eg1c.out", "w"))
# print(fit)
expval = lsqfit.BayesIntegrator(fit, limit=10.)
# adapt integrator to PDF
expval(neval=40000, nitn=10)
# calculate expectation value of function g(p)
fit_hist = gv.PDFHistogram(fit.p['E'][0])
def g(p):
parameters = [p['a'][0], p['E'][0]]
return dict(
mean=parameters,
outer=np.outer(parameters, parameters),
hist=fit_hist.count(p['E'][0]),
)
r = expval(g, neval=40000, nitn=10, adapt=False)
# print(results)
print(r.summary())
means = r['mean']
cov = r['outer'] - np.outer(r['mean'], r['mean'])
print('Results from Bayesian Integration:')
print('a0: mean =', means[0], ' sdev =', cov[0, 0]**0.5)
print('E0: mean =', means[1], ' sdev =', cov[1, 1]**0.5)
print('covariance from Bayesian integral =',
np.array2string(cov, prefix=36 * ' '))
print
print('Results from Least-Squares Fit:')
print('a0: mean =', fit.p['a'][0].mean, ' sdev =', fit.p['a'][0].sdev)
print('E0: mean =', fit.p['E'][0].mean, ' sdev =', fit.p['E'][0].sdev)
print(
'covariance from least-squares fit =',
np.array2string(gv.evalcov([fit.p['a'][0], fit.p['E'][0]]),
prefix=36 * ' ',
precision=3))
sys.stdout = sys_stdout
# make histogram of E[0] probabilty
plt = fit_hist.make_plot(r['hist'])
plt.xlabel('$E_0$')
plt.ylabel('probability')
plt.savefig('eg1c.png', bbox_inches='tight')
# plt.show()
if DO_BOOTSTRAP:
Nbs = 40 # number of bootstrap copies
outputs = {
'E1/E0': [],
'E2/E0': [],
'a1/a0': [],
'a2/a0': [],
'E1': [],
'a1': []
} # results
for bsfit in fit.bootstrap_iter(n=Nbs):
E = bsfit.pmean['E'] # best-fit parameters
a = bsfit.pmean['a']
outputs['E1/E0'].append(E[1] / E[0]) # accumulate results
outputs['E2/E0'].append(E[2] / E[0])
outputs['a1/a0'].append(a[1] / a[0])
outputs['a2/a0'].append(a[2] / a[0])
outputs['E1'].append(E[1])
outputs['a1'].append(a[1])
# print(E[:2])
# print(a[:2])
# print(bsfit.chi2/bsfit.dof)
# extract means and standard deviations from the bootstrap output
for k in outputs:
outputs[k] = gv.gvar(np.mean(outputs[k]), np.std(outputs[k]))
print('Bootstrap results:')
print('E1/E0 =', outputs['E1/E0'], ' E2/E1 =', outputs['E2/E0'])
print('a1/a0 =', outputs['a1/a0'], ' a2/a0 =', outputs['a2/a0'])
print('E1 =', outputs['E1'], ' a1 =', outputs['a1'])
}
# a priori values for fit parameters
prior = dict(a=gv.gvar(0.5, 0.5), b=gv.gvar(0.5, 0.5))
# print(y["data1"][0].mean,"+-",y["data1"][0].sdev)
# print(gv.evalcov(y["data1"]))
def fcn(x, p): # fit function of x and parameters p
ans = {}
for k in ["data1", "data2"]:
ans[k] = gv.exp(p['a'] + x[k] * p['b'])
ans['b/a'] = p['b'] / p['a']
return ans
# do the fit
fit = lsqfit.nonlinear_fit(data=(x, y), prior=prior, fcn=fcn)
sys.stdout = open("eg0.out", "w")
print(fit.format(maxline=True)) # print standard summary of fit
p = fit.p # best-fit values for parameters
outputs = dict(a=p['a'], b=p['b'])
outputs['b/a'] = p['b'] / p['a']
inputs = dict(y=y, prior=prior)
print(gv.fmt_values(outputs)) # tabulate outputs
print(gv.fmt_errorbudget(outputs, inputs)) # print error budget for outputs
# save best-fit values in file "outputfile.p" for later use
import pickle
pickle.dump(fit.p, open("outputfile.p", "wb"))
def _get_error_budget(self, verbose=False, **kwargs):
output = None
for observable in ['w0', 't0']:
# Fill these out
disc_keys = [
'A_a', 'A_alpha', 'A_aa', 'A_al', 'A_as', 'A_aaa', 'A_aal',
'A_aas', 'A_all', 'A_als', 'A_ass'
]
chiral_keys = [
'c0', 'A_l', 'A_s', 'A_ll', 'A_ls', 'A_ss', 'A_ll_g', 'A_lll',
'A_lls', 'A_lss', 'A_sss', 'A_lll_g', 'A_lls_g', 'A_lll_gg'
]
phys_keys = list(self.phys_point_data)
stat_key = 'lam_chi' # Since the input data is correlated, only need a single variable as a proxy for all
if verbose:
if output is None:
output = ''
inputs = {}
# xpt/chiral contributions
inputs.update({
str(param) + ' [disc]': self.prior[observable][param]
for param in disc_keys if param in self.prior[observable]
})
inputs.update({
str(param) + ' [xpt]': self.prior[observable][param]
for param in chiral_keys if param in self.prior[observable]
})
# phys point contributions
inputs.update({
str(param) + ' [pp]': self.phys_point_data[param]
for param in list(phys_keys)
})
del inputs['lam_chi [pp]']
# stat contribtions
inputs.update({
'x [stat]': self._get_prior(stat_key)[observable],
'y [stat]': self.fitter[observable].y
})
if kwargs is None:
kwargs = {}
kwargs.setdefault('percent', False)
kwargs.setdefault('ndecimal', 10)
kwargs.setdefault('verify', True)
if observable == 'w0':
output += 'observable: ' + observable + '\n' + gv.fmt_errorbudget(
outputs={'w0': self.w0
}, inputs=inputs, **kwargs) + '\n---\n'
elif observable == 't0':
output += 'observable: ' + observable + '\n' + gv.fmt_errorbudget(
outputs={'t0': self.sqrt_t0
}, inputs=inputs, **kwargs) + '\n---\n'
else:
if output is None:
output = {}
output[observable] = {}
if observable == 'w0':
value = self.w0
elif observable == 't0':
value = self.sqrt_t0
output[observable]['disc'] = value.partialsdev([
self.prior[observable][key] for key in disc_keys
if key in self.prior[observable]
])
output[observable]['chiral'] = value.partialsdev([
self.prior[observable][key] for key in chiral_keys
if key in self.prior[observable]
])
output[observable]['pp_input'] = value.partialsdev(
[self.phys_point_data[key] for key in phys_keys])
output[observable]['stat'] = value.partialsdev(
[
self._get_prior(stat_key)[observable],
self.fitter[observable].y
]
#self.fitter['w0'].y
)
return output
def do_fit(svdcut=None, do_plot=False):
if svdcut is None:
svdcut = lsqfit.nonlinear_fit.set()['svdcut']
sys.stdout = tee.tee(sys_stdout, open('eg5a.out', 'w'))
default_svd = True
else:
default_svd = False
x, y = make_data()
prior = make_prior(
100) # 20 exponential terms in all (10 gives same result)
p0 = None
for nexp in range(1, 6):
# marginalize the last 100 - nexp terms
fit_prior = gv.BufferDict() # part of prior used in fit
ymod_prior = gv.BufferDict() # part of prior absorbed in ymod
for k in prior:
fit_prior[k] = prior[k][:nexp]
ymod_prior[k] = prior[k][nexp:]
ymod = y - fcn(x, ymod_prior)
# fit modified data with just nexp terms
fit = lsqfit.nonlinear_fit(data=(x, ymod),
prior=fit_prior,
fcn=fcn,
p0=p0,
tol=1e-10,
svdcut=svdcut)
if not default_svd and nexp == 5:
sys.stdout = tee.tee(sys_stdout, open('eg5b.out', 'w'))
print '************************************* nexp =', nexp
print fit.format(True)
p0 = fit.pmean
if do_plot:
import matplotlib.pyplot as plt
if nexp > 4:
continue
plt.subplot(2, 2, nexp)
if nexp not in [1, 3]:
plt.yticks([0.05, 0.10, 0.15, 0.20, 0.25], [])
else:
plt.ylabel('y')
if nexp not in [3, 4]:
plt.xticks([1.0, 1.5, 2.0, 2.5], [])
else:
plt.xlabel('x')
plt.errorbar(x=x, y=gv.mean(ymod), yerr=gv.sdev(ymod), fmt='bo')
plt.plot(x, y, '-r')
plt.plot(x, fcn(x, fit.pmean), ':k')
plt.text(1.75, 0.22, 'nexp = {}'.format(nexp))
if nexp == 4:
plt.savefig('eg5.png', bbox_inches='tight')
plt.show()
# print summary information and error budget
E = fit.p['E'] # best-fit parameters
a = fit.p['a']
outputs = {
'E1/E0': E[1] / E[0],
'E2/E0': E[2] / E[0],
'a1/a0': a[1] / a[0],
'a2/a0': a[2] / a[0]
}
inputs = {
'E prior': prior['E'],
'a prior': prior['a'],
'svd cut': fit.svdcorrection,
}
print(gv.fmt_values(outputs))
print(gv.fmt_errorbudget(outputs, inputs))
sys.stdout = sys_stdout
