def animate(i):
gen = gg.GaussDistribution((2*i) + 1, sigma=sigma).sample(nentries=stats)
x, y = plotter.getplotxy(gen)
line.set_xdata(x) # update the data
line.set_ydata(y) # update the data
return line,
import gaussguess as gg
import gaussguess.plotter as plotter
NBINS = 21
sigma = 0.15
fig, ax = plt.subplots()
classes = ["gauss", "uniform", "triangular", "laplace", "poisson"]
iclasses = range(len(classes))
classifier = gg.DistributionMultiLabelClassifier(NBINS, nlabels=len(classes))
classifier.loadmodel("{}bins_model_multilabel.h5".format(NBINS))
normop = lambda x: x / max(x) if max(x) > 0 else 0
dist = gg.GaussDistribution(NBINS, sigma=sigma, xlim=[0, 1])
gen = dist.sample(nentries=1).normalise()
prediction = classifier.predict(gen)
x, y = plotter.getplotxy(gen.normalise(op=normop))
lims = gen.limits
line, = ax.plot(x, y, 'k', alpha=0.6)
stat_label = ax.text(0.01, 0.99, "{} entries".format(1))
texts = []
startX = 0.7
startY = 0.95
for i, c in enumerate(classes):
texts.append(
ax.text(startX, startY, "{} = {:.2f}%".format(c, prediction[i] * 100)))
startY -= 0.05
import matplotlib.pyplot as plt
import matplotlib.animation as animation
import numpy as np
import gaussguess as gg
import gaussguess.plotter as plotter
stats = 100000
sigma = 0.05
fig, ax = plt.subplots()
gen = gg.GaussDistribution(3, sigma=sigma).sample(nentries=stats)
x, y = plotter.getplotxy(gen)
lims = gen.limits
line, = ax.plot(x, y, 'k', alpha=0.6)
def animate(i):
gen = gg.GaussDistribution((2*i) + 1, sigma=sigma).sample(nentries=stats)
x, y = plotter.getplotxy(gen)
line.set_xdata(x) # update the data
line.set_ydata(y) # update the data
return line,
ani = animation.FuncAnimation(fig, animate, np.arange(0, 100),
interval=50, blit=False)
plt.xlim(lims)
plt.ylabel("count / {:.3e}".format(gen.binwidth))
plt.title("Increasing number of bins")
import gaussguess as gg nbins = 9 stats = 100 sn_ratio = 100 signal = gg.GaussDistribution(nbins, sigma=0.1).sample(nentries=stats) noise = gg.UniformDistribution(nbins).sample(nentries=100) measured = signal + (stats / sn_ratio) * noise # print(gen.values) gg.histplot(signal).show() gg.histplot(noise).show() gg.histplot(measured).show()
import gaussguess as gg
import gaussguess.plotter as plotter
nbins = 11
stats = 1000
dists = [
gg.GaussDistribution(nbins, sigma=0.1),
gg.GaussDistribution(nbins, sigma=1.0),
gg.GaussDistribution(nbins, sigma=2.0, xlim=[0, 10]),
gg.LaplaceDistribution(nbins, lam=0.1),
gg.LaplaceDistribution(nbins, lam=1.2),
gg.UniformDistribution(nbins),
gg.UniformDistribution(3),
gg.TriangularDistribution(nbins),
gg.TriangularDistribution(nbins, center=0.4),
gg.TriangularDistribution(47),
gg.PoissonDistribution(nbins),
gg.PoissonDistribution(nbins, lam=6, xlim=[0, 10]),
]
for d in dists:
result = d.sample(nentries=stats)
plotter.histplot(result).show()
import gaussguess as gg
NBINS = 21
signaldists = [
gg.GaussDistribution(NBINS, sigma=0.05),
gg.GaussDistribution(NBINS, sigma=0.1),
gg.GaussDistribution(NBINS, sigma=0.15),
gg.GaussDistribution(NBINS, sigma=0.2),
gg.GaussDistribution(NBINS, sigma=0.25),
gg.GaussDistribution(NBINS, sigma=0.3),
gg.GaussDistribution(NBINS, sigma=0.4),
gg.GaussDistribution(NBINS, sigma=2.0, xlim=[0, 10]),
]
backgrounddists = [
gg.LaplaceDistribution(NBINS, lam=0.1),
gg.LaplaceDistribution(NBINS, lam=1.2),
gg.UniformDistribution(NBINS),
gg.TriangularDistribution(NBINS),
gg.TriangularDistribution(NBINS, center=0.4),
gg.TriangularDistribution(NBINS, center=0.3),
gg.TriangularDistribution(NBINS, center=0.8),
gg.PoissonDistribution(NBINS),
gg.PoissonDistribution(NBINS, lam=6, xlim=[0, 10]),
]
classifier = gg.DistributionBinaryClassifier(NBINS)
classifier.generatedata(signaldists,
backgrounddists,
nloops=50000,
import gaussguess as gg
NTRUE = 100000
NFALSE = 100000
NBINS = 5
SIGMA = 1.0
STATS = 100
counter = 0
with open('samples_size{}.csv'.format(NBINS), 'wt') as csv:
csv.write("index,sigma,stats,isgauss,{}\n".format(",".join(
["bin{}".format(i) for i in range(NBINS)])))
for _ in range(NTRUE):
signal = gg.GaussDistribution(NBINS,
sigma=SIGMA).sample(nentries=STATS)
csv.write("{},{},{},{}".format(counter, SIGMA, STATS, 1))
for v in signal.values:
csv.write(",{:.15e}".format(v))
csv.write("\n")
counter += 1
for _ in range(NTRUE):
background = gg.UniformDistribution(NBINS).sample(nentries=STATS)
csv.write("{},{},{},{}".format(counter, SIGMA, STATS, 0))
for v in background.values:
csv.write(",{:.15e}".format(v))
csv.write("\n")
counter += 1
import gaussguess as gg
NBINS = 111
classes = ["gauss", "uniform", "triangular", "laplace", "poisson"]
iclasses = range(len(classes))
getlabel = lambda name: iclasses[classes.index(name)]
labeled_data = [
(getlabel("gauss"), gg.GaussDistribution(NBINS, sigma=0.05)),
(getlabel("gauss"), gg.GaussDistribution(NBINS, sigma=0.10)),
(getlabel("gauss"), gg.GaussDistribution(NBINS, sigma=0.15)),
(getlabel("gauss"), gg.GaussDistribution(NBINS, sigma=0.75)),
(getlabel("gauss"), gg.GaussDistribution(NBINS, sigma=0.20)),
(getlabel("gauss"), gg.GaussDistribution(NBINS, sigma=0.20)),
(getlabel("laplace"), gg.LaplaceDistribution(NBINS, lam=0.1)),
(getlabel("laplace"), gg.LaplaceDistribution(NBINS, lam=0.12)),
(getlabel("laplace"), gg.LaplaceDistribution(NBINS, lam=0.052)),
(getlabel("uniform"), gg.UniformDistribution(NBINS)),
(getlabel("uniform"), gg.UniformDistribution(NBINS)),
(getlabel("uniform"), gg.UniformDistribution(NBINS)),
(getlabel("uniform"), gg.UniformDistribution(NBINS)),
(getlabel("triangular"), gg.TriangularDistribution(NBINS, center=0.1)),
(getlabel("triangular"), gg.TriangularDistribution(NBINS, center=0.2)),
(getlabel("triangular"), gg.TriangularDistribution(NBINS, center=0.5)),
(getlabel("triangular"), gg.TriangularDistribution(NBINS, center=0.6)),
(getlabel("triangular"), gg.TriangularDistribution(NBINS, center=0.99)),
(getlabel("triangular"), gg.TriangularDistribution(NBINS, center=0.4)),
(getlabel("poisson"), gg.PoissonDistribution(NBINS, lam=6, xlim=[0, 10])),
(getlabel("poisson"), gg.PoissonDistribution(NBINS, lam=2, xlim=[0, 10])),
(getlabel("poisson"), gg.PoissonDistribution(NBINS, lam=3, xlim=[0, 10])),