def test_extended_ulh(self):
eg = Extended(gaussian)
lh = UnbinnedLH(eg, self.data, extended=True, extended_bound=(-20, 20))
minuit = Minuit(lh, mean=4.5, sigma=1.5, N=19000.,
pedantic=False, print_level=0)
minuit.migrad()
assert_allclose(minuit.values['N'], 20000, atol=sqrt(20000.))
assert minuit.migrad_ok()
def test_extended_ulh_2(self):
eg = Extended(gaussian)
lh = UnbinnedLH(eg, self.data, extended=True)
minuit = Minuit(lh,
mean=4.5,
sigma=1.5,
N=19000.,
pedantic=False,
print_level=0)
minuit.migrad()
assert (minuit.migrad_ok())
assert_almost_equal(minuit.values['N'], 20000, delta=sqrt(20000.))
from iminuit import Minuit
from probfit import UnbinnedLH, gaussian, Extended
from matplotlib import pyplot as plt
from numpy.random import randn
data = randn(1000) * 2 + 1
ulh = UnbinnedLH(gaussian, data)
m = Minuit(ulh, mean=0., sigma=0.5)
plt.figure(figsize=(8, 6))
plt.subplot(221)
ulh.draw(m)
plt.title('Unextended Before')
m.migrad() # fit
plt.subplot(222)
ulh.draw(m)
plt.title('Unextended After')
#Extended
data = randn(2000) * 2 + 1
egauss = Extended(gaussian)
ulh = UnbinnedLH(egauss, data, extended=True, extended_bound=(-10., 10.))
m = Minuit(ulh, mean=0., sigma=0.5, N=1800.)
plt.subplot(223)
ulh.draw(m)
plt.title('Extended Before')
# <codecell>
#you can define your pdf manually like this
def my_gauss(x, mu, sigma):
return exp(-0.5*(x-mu)**2/sigma**2)/(sqrt(2*pi)*sigma)
# <codecell>
#probfit use the same describe magic as iminuit
#Build your favorite cost function like this
#Notice no RooRealVar etc. It use introspection to
#find parameters
#the only requirement is that the first argument is
#in independent variable the rest are parameters
ulh = UnbinnedLH(my_gauss, gdata)
describe(ulh)
# <codecell>
m = Minuit(ulh, mu=0.2, sigma=1.5)
m.set_up(0.5) #remember up is 0.5 for likelihood and 1 for chi^2
ulh.show(m)
# <codecell>
m.migrad();
ulh.show(m)
# <markdowncell>
# <codecell>
#you can define your pdf manually like this
def my_gauss(x, mu, sigma):
return exp(-0.5*(x-mu)**2/sigma**2)/(sqrt(2*pi)*sigma)
# <codecell>
#probfit use the same describe magic as iminuit
#Build your favorite cost function like this
#Notice no RooRealVar etc. It use introspection to
#find parameters
#the only requirement is that the first argument is
#in independent variable the rest are parameters
ulh = UnbinnedLH(my_gauss, gdata)
describe(ulh)
# <codecell>
m = Minuit(ulh, mu=0.2, sigma=1.5)
m.set_up(0.5) #remember up is 0.5 for likelihood and 1 for chi^2
ulh.show(m)
# <codecell>
m.migrad();
ulh.show(m)
# <markdowncell>
pdf= AddPdfNorm(g0,g1) describe(pdf) # <codecell> seed(0) toydata = gen_toy(pdf, 1000,(-10,10), m0=-2, m1=2, s0=1, s1=1, f_0=0.3, quiet=False) # <codecell> inipars= dict(m0=0, m1=0, s0=1, s1=1, f_0=0.5, error_m0=0.1, error_m1=0.1, error_s0=0.1, error_s1=0.1, error_f_0=0.1) # <codecell> # Normal fit uh1= UnbinnedLH(pdf, toydata) m1= Minuit(uh1, print_level=1, **inipars) m1.migrad(); uh1.draw(); print m1.values # <codecell> # Blind one parameter uh2= UnbinnedLH( BlindFunc(pdf, toblind='m1', seedstring='some_random_stuff', width=0.5, signflip=False), toydata) m2= Minuit(uh2, print_level=1, **inipars) m2.migrad(); uh2.draw(); print m2.values # <codecell>
# ###Fitting # <codecell> from iminuit import Minuit from probfit import UnbinnedLH, gaussian # <codecell> data = randn(10000) hist(data, bins=100, histtype='step') # <codecell> ulh = UnbinnedLH(gaussian, data) ulh.draw(args=dict(mean=1.2, sigma=0.7)) # <codecell> m = Minuit(ulh, mean=1.2, sigma=0.7) # <codecell> m.migrad() # <codecell> print m.values print m.errors ulh.draw(m)
def model(fit_range, bin):
nrm_bkg_pdf = Normalized(rename(exp, ['x', 'l%d' % bin]), fit_range)
ext_bkg_pdf = Extended(nrm_bkg_pdf, extname='Ncomb_%d' % bin)
ext_sig_pdf = Extended(rename(gaussian, ['x', 'm%d' % bin, "sigma%d" % bin]), extname='Nsig_%d' % bin)
tot_pdf = AddPdf(ext_bkg_pdf, ext_sig_pdf)
print('pdf: {}'.format(describe(tot_pdf)))
return tot_pdf
fit_range = (2900, 3300)
mod_1 = model(fit_range, 1)
lik_1 = UnbinnedLH(mod_1, tot_m, extended=True)
mod_2 = model(fit_range, 2)
lik_2 = UnbinnedLH(mod_2, tot_u, extended=True)
sim_lik = SimultaneousFit(lik_1, lik_2)
describe(sim_lik)
pars = dict(l1=0.002, Ncomb_1=1000, m1=3100, sigma1=10, Nsig_1=1000, l2=0.002, Ncomb_2=1000, m2=3100, sigma2=10,
Nsig_2=1000)
minuit = Minuit(sim_lik, pedantic=False, print_level=0, **pars)
# In[8]:
if do_probfit:
start = time.time()
minuit.migrad()
from iminuit import Minuit
from probfit import UnbinnedLH, gaussian, SimultaneousFit, rename
from matplotlib import pyplot as plt
from numpy.random import randn, seed
seed(0)
width = 2.
data1 = randn(1000) * width + 1
data2 = randn(1000) * width + 2
#two gaussian with shared width
pdf1 = rename(gaussian, ('x', 'mu_1', 'sigma'))
pdf2 = rename(gaussian, ('x', 'mu_2', 'sigma'))
lh1 = UnbinnedLH(pdf1, data1)
lh2 = UnbinnedLH(pdf2, data2)
simlh = SimultaneousFit(lh1, lh2)
m = Minuit(simlh, mu_1=1.2, mu_2=2.2, sigma=1.5)
plt.figure(figsize=(8, 3))
plt.subplot(211)
simlh.draw(m)
plt.suptitle('Before')
m.migrad() # fit
plt.figure(figsize=(8, 3))
plt.subplot(212)
simlh.draw(m)
from iminuit import Minuit
from probfit import UnbinnedLH, gaussian, Extended
from matplotlib import pyplot as plt
from numpy.random import randn
data = randn(1000)*2 + 1
ulh = UnbinnedLH(gaussian, data)
m = Minuit(ulh, mean=0., sigma=0.5)
plt.figure(figsize=(8, 6))
plt.subplot(221)
ulh.draw(m)
plt.title('Unextended Before')
m.migrad() # fit
plt.subplot(222)
ulh.draw(m)
plt.title('Unextended After')
#Extended
data = randn(2000)*2 + 1
egauss = Extended(gaussian)
ulh = UnbinnedLH(egauss, data, extended=True, extended_bound=(-10.,10.))
m = Minuit(ulh, mean=0., sigma=0.5, N=1800.)
plt.subplot(223)
ulh.draw(m)
plt.title('Extended Before')
