def ent(k, x, y):
print("Entropía de X:")
print("Estimador 1: ", mi.entropy(x, k=k))
print("Estimador 2: ", ee.entropy(x))
print("Entropía de Y:")
print("Estimador 1: ", mi.entropy(y, k=k))
print("Estimador 2: ", ee.entropy(y))
def computeEntropies(self):
for var in self.data.getSeriesNames():
d = self.data.getSeries(var)
#print('p1')
p = self.mi_prepare(d)
#print('p2')
entropy = ee.entropy(p)
self.entropies[var] = entropy
Dprint("Entropy of ", var, "=", entropy)
return
def allAverageSubsetEntropies(variables):
entropies = []
a = 1
while a < len(variables):
subsets = randomSubsets(len(variables), a)
entropy1 = averageSubsetEntropy(variables, subsets)
entropies.append(entropy1)
a += 1
bigSetEntropy = ee.entropy(list(np.array(variables).T), k=10)
return (entropies, bigSetEntropy)
def example_ent_unif(dat, k=3):
x = dat.X
a = dat.a
b = dat.b
H_est = mi.entropy(x, k)
H_est2 = ee.entropy(x, k)
print("Entropía x:")
print("Teórica: ", np.log(b - a) / np.log(2))
print("Estimador 1: ", H_est)
print("Estimador 2: ", H_est2)
def averageSubsetEntropy(variables, subsets):
totalEntropy = 0.0
a = 0
while a < len(subsets):
subset = np.array(subsets[a])
vars1 = np.array(variables)[subset]
vars1 = list(vars1.T)
entropy = ee.entropy(vars1, k=10)
totalEntropy += entropy
a += 1
totalEntropy = totalEntropy / float(len(subsets))
return totalEntropy
def example_ent():
x = np.random.choice([1.0, 2.0, 3.0, 4.0], (1000, 1),
p=[0.5, 0.25, 0.125, 0.125])
H_est = mi.entropy(x, k=5)
H_est2 = ee.entropy(x, k=5)
print("Entropía x:")
print("Teórica: ", 7.0 / 4.0)
print("Estimador 1: ", H_est)
print("Estimador 2: ", H_est2)
y = np.random.choice([1.0, 2.0, 3.0, 4.0], (1000, 1),
p=[0.25, 0.25, 0.25, 0.25])
H_est = mi.entropy(y, k=5)
H_est2 = ee.entropy(y, k=5)
print("Entropía y:")
print("Teórica: ", 2.0)
print("Estimador 1: ", H_est)
print("Estimador 2: ", H_est2)
def test_entropy(dat, k=3):
C = dat.C
X = dat.X
H_th = mi.entropy_gaussian(C) / np.log(2)
H_est = mi.entropy(X, k)
H_est2 = ee.entropy(X, k)
print("Entropía gaussiana:")
print("Teórica: ", H_th)
print("Estimador 1: ", H_est)
print("Estimador 2: ", H_est2)
# La entropía estimada debe ser menor que la teórica pero no por mucho
np.testing.assert_array_less(H_est, H_th)
np.testing.assert_array_less(H_est2, H_th)
np.testing.assert_array_less(.9 * H_th, H_est)
np.testing.assert_array_less(.9 * H_th, H_est2)
def ttest(N, ds, ns, k, save):
# DataFrame - MultiIndex para almacenar mediciones
# Creamos nombres de filas y columnas
iterables = [ds, ['p', 't']]
index = pd.MultiIndex.from_product(iterables, names=['ds', 'test-res'])
iterables2 = [['ent', 'mi'], ns]
cols = pd.MultiIndex.from_product(iterables2, names=['función', 'ns'])
# Creamos el DataFrame
df = pd.DataFrame(index=index, columns=cols)
for d in ds:
for n in ns:
# ENTROPÍA
# Generamos los datos
X = [data.Data('random', n, d).X for i in range(0, N)]
# Calculamos la entropía
mi_ent = np.array([mi.entropy(x, k=k) for x in X])
ee_ent = np.array([ee.entropy(x, k=k) for x in X])
# Calculamos el ttest
t, p = stats.ttest_rel(mi_ent, ee_ent)
# Almacenamos los datos
df.loc[(d, 'p'), ('ent', n)] = p
df.loc[(d, 't'), ('ent', n)] = t
# INFORMACIÓN MUTUA
Y = [data.Data('random', n, d).X for i in range(0, N)]
mi_mi = np.array(
[mi.mutual_information((x, y), k=k) for x, y in zip(X, Y)])
ee_mi = np.array([ee.mi(x, y, k=k) for x, y in zip(X, Y)])
# Calculamos el ttest
t, p = stats.ttest_rel(mi_mi, ee_mi)
# Almacenamos los datos
df.loc[(d, 'p'), ('mi', n)] = p
df.loc[(d, 't'), ('mi', n)] = t
# Imprimimos en un archivo por cada función los resultados hasta el momento
if (save):
df.to_pickle("./stats/ttest.pkl")
def test_normality(N, ds, ns, k):
for d in ds:
for n in ns:
print("ENTROPÍA")
# Generamos los datos
X = np.array([data.Data('random', n, d).X for i in range(0, N)])
# Calculamos la entropía
mi_ent = np.array([mi.entropy(x, k=k) for x in X])
ee_ent = np.array([ee.entropy(x, k=k) for x in X])
# Calculamos las diferencias entre las entropías
D = np.array([X1 - X2 for X1, X2 in zip(mi_ent, ee_ent)])
# Realizamos el test de normalidad
k2, p = stats.normaltest(D)
print("d: ", d, ", n: ", n)
if p <= 0.05: # Hipótesis nula: D proviene de una distribución normal
print("Podemos rechazar la hipótesis nula")
else:
print("No podemos rechazar la hipótesis nula")
print("INFORMACIÓN MUTUA")
# Generamos los datos
Y = [data.Data('random', n, d).X for i in range(0, N)]
# Calculamos la información mutua
mi_mi = np.array(
[mi.mutual_information((x, y), k=k) for x, y in zip(X, Y)])
ee_mi = np.array([ee.mi(x, y, k=k) for x, y in zip(X, Y)])
# Calculamos las diferencias entre las implementaciones
D = np.array([X1 - X2 for X1, X2 in zip(mi_mi, ee_mi)])
# Realizamos el test de normalidad
k2, p = stats.normaltest(D)
print("d: ", d, ", n: ", n)
if p <= 0.05: # Hipótesis nula: D proviene de una distribución normal
print("Podemos rechazar la hipótesis nula")
else:
print("No podemos rechazar la hipótesis nula")
def err_entropy(dat, k=3, show=False):
C = dat.C
X = dat.X
# Calculamos las entropías
H_th = mi.entropy_gaussian(C) / np.log(2)
H_est = mi.entropy(X, k)
H_est2 = ee.entropy(X, k)
if (show):
print("Entropía gaussiana:")
print("Teórica: ", H_th)
print("Estimador 1: ", H_est)
print("Estimador 2: ", H_est2)
# Calculamos la diferencia entre la entropía teórica y la estimada
dif1 = abs(H_th - H_est)
dif2 = abs(H_th - H_est2)
return dif1, dif2
def computeEntropies(self):
eList = []
eList2 = []
for var in self.data.getSeriesNames():
d = self.data.getSeries(var)
#print('len = ', len(d))
#print('p1')
p = self.mi_prepare(d)
#print('p2')
entropy = ee.entropy(p, base=math.e)
self.entropies[var] = entropy
dentropy = self.dentropy(var)
#print("Entropy, dentropy of ", var, "=", entropy, dentropy)
eList.append((dentropy, var))
eList2.append((entropy, var))
eList.sort()
#print('eList = ', eList)
eList2.sort()
#print('eList2 = ', eList2)
return
def test_entropy():
np.random.seed(12345)
results = [[ee.entropy(np.random.rand(200, 2), k=j) for j in range(1, 6)]
for _ in range(200)]
assert (np.all(np.diff(np.mean(results, axis=0)) > 0))
#!/bin/env python
# Testing the NPEET estimators
import entropy_estimators as ee
from math import log, pi
import numpy as np
import numpy.random as nr
import random
from numpy.linalg import det
# Some test cases to see usage and correctness
# Differential entropy estimator
print("For a uniform distribution with width alpha, the differential entropy is log_2 alpha, setting alpha = 2")
print("and using k=1, 2, 3, 4, 5")
print("result:", [ee.entropy([[2 * random.random()] for i in range(1000)], k=j + 1) for j in range(5)])
# CONDITIONAL MUTUAL INFORMATION
Ntry = [10000] #Number of samples to use in estimate
nsamples = 5 # Number of times to est mutual information for CI
samplo = int(0.025 * nsamples) # confidence intervals
samphi = int(0.975 * nsamples)
# print('\nGaussian random variables\n')
# print('Conditional Mutual Information')
d1 = [1, 1, 0]
d2 = [1, 0, 1]
d3 = [0, 1, 1]
mat = [d1, d2, d3]
tmat = np.transpose(mat)
diag = [[3, 0, 0], [0, 1, 0], [0, 0, 1]]
import entropy_estimators as ee
from math import log, pi
import numpy as np
import numpy.random as nr
import random
from numpy.linalg import det
# Some test cases to see usage and correctness
# Differential entropy estimator
print(
"For a uniform distribution with width alpha, the differential entropy is log_2 alpha, setting alpha = 2"
)
print("and using k=1, 2, 3, 4, 5")
print("result:", [
ee.entropy([[2 * random.random()] for i in range(1000)], k=j + 1)
for j in range(5)
])
# CONDITIONAL MUTUAL INFORMATION
Ntry = [10, 25, 50, 100,
200] # , 1000, 2000] #Number of samples to use in estimate
nsamples = 100 # Number of times to est mutual information for CI
samplo = int(0.025 * nsamples) # confidence intervals
samphi = int(0.975 * nsamples)
print('\nGaussian random variables\n')
print('Conditional Mutual Information')
d1 = [1, 1, 0]
d2 = [1, 0, 1]
d3 = [0, 1, 1]
#python2.7 #Testing the NPEET estimators import entropy_estimators as ee from math import log,pi import numpy as np import numpy.random as nr import random from numpy.linalg import det #Some test cases to see usage and correctness ## Differential entropy estimator print "For a uniform distribution with width alpha, the differential entropy is log_2 alpha, setting alpha = 2" print "and using k=1,2,3,4,5" print "result:", [ee.entropy([[2*random.random()] for i in range(1000)],k=j+1) for j in range(5)] ## CONDITIONAL MUTUAL INFORMATION Ntry = [10,25,50,100,200] #,1000,2000] #Number of samples to use in estimate nsamples = 100 #Number of times to est mutual information for CI samplo = int(0.025*nsamples) #confidence intervals samphi = int(0.975*nsamples) print '\nGaussian random variables\n' print 'Conditional Mutual Information' d1 = [1,1,0] d2 = [1,0,1] d3 = [0,1,1] mat = [d1,d2,d3] tmat = np.transpose(mat) diag = [[3,0,0],[0,1,0],[0,0,1]]
def calculateEntropy(x, k):
return ee.entropy(np.array([x]).T, k=k)
def logging_entropy(log_path):
if os.path.isfile(os.path.join(log_path, 'entropy.txt')):
return
img = cv2.imread(
'/home/alberto/catkin_ws/src/pf2d_localizer/worlds/test_environment_similarities/test_environment_similarities.png',
0)
with open(
'/home/alberto/DATA/Dropbox/work/catkin_src/pf2d_localizer/worlds/test_environment_similarities/test_environment_similarities.yaml',
'r') as stream:
yaml_data = yaml.load(stream)
x_min = -img.shape[1] / 2.0 * float(yaml_data['resolution'])
y_max = img.shape[0] / 2.0 * float(yaml_data['resolution'])
onlyfiles = [
f for f in os.listdir(log_path)
if os.path.isfile(os.path.join(log_path, f)) and "particles" in f
and "txt" in f
]
onlyfiles = sorted(
onlyfiles, key=lambda f: int(os.path.splitext(f)[0].split('_')[1]))
entropies = []
for f in onlyfiles:
#print onlyfiles
particles = numpy.genfromtxt(os.path.join(log_path, f),
unpack=True)
#img = cv2.imread('/home/alberto/DATA/Dropbox/work/catkin_src/pf2d_localizer/worlds/map_turtlebot/map_turtlebot.pgm', 0)
"""
total_weight = 0.0
for w in particles[3]:
total_weight += w
for w in particles[3]:
#print "prev ", w
w /= total_weight
#print "aft ", w
bins = numpy.add.accumulate(particles[3])
values_indices = numpy.digitize(numpy.random.random_sample(10000), bins)"""
"""for v in values_indices:
if particles[3][v] == 0:
print particles[3][v]"""
if consider_angle:
#particles_pair = [list(p) for p in zip(particles[0]-x_min/float(yaml_data['resolution']), (y_max-particles[1])/float(yaml_data['resolution']), particles[2]/3.14*180+180)] # TODO weight
particles_pair = [
list(p)
for p in zip(particles[0], particles[1], particles[2])
] # TODO weight
#particles_pair = [list([particles[0][v]-x_min/float(yaml_data['resolution']), (y_max-particles[1][v])/float(yaml_data['resolution']), particles[2][v]/3.14*180+180]) for v in values_indices if particles[3][v] > 0.000001] # TODO weight
else:
#particles_pair = [list(p) for p in zip(particles[0]-x_min/float(yaml_data['resolution']), (y_max-particles[1])/float(yaml_data['resolution']))]
particles_pair = [
list(p) for p in zip(particles[0], particles[1])
] # TODO weight
#particles_pair = [list([particles[0][v-1]-x_min/float(yaml_data['resolution']), (y_max-particles[1][v-1])/float(yaml_data['resolution'])]) for v in values_indices if particles[3][v-1] > 0.0001]
entropies.append(ee.entropy(particles_pair))
with open(os.path.join(log_path, 'entropy.txt'),
'w') as f: # TODO parameter
for e in entropies:
f.write(str(e) + '\n')