def RunGMMShogun(q):
totalTimer = Timer()
try:
# Load input dataset.
Log.Info("Loading dataset", self.verbose)
dataPoints = np.genfromtxt(self.dataset, delimiter=',')
dataFeat = RealFeatures(dataPoints.T)
# Get all the parameters.
g = re.search("-g (\d+)", options)
n = re.search("-n (\d+)", options)
s = re.search("-n (\d+)", options)
g = 1 if not g else int(g.group(1))
n = 250 if not n else int(n.group(1))
# Create the Gaussian Mixture Model.
model = SGMM(g)
model.set_features(dataFeat)
with totalTimer:
model.train_em(1e-9, n, 1e-9)
except Exception as e:
q.put(-1)
return -1
time = totalTimer.ElapsedTime()
q.put(time)
return time
def fit_gmm(self, samples):
"""
Runs a couple of em instances on random starting points and returns
internal GMM representation of best instance
"""
features = RealFeatures(samples.T)
gmms = []
log_likelihoods = zeros(self.num_runs_em)
for i in range(self.num_runs_em):
# set up Shogun's GMM class and run em (corresponds to random
# initialisation)
gmm = GMM(self.num_components)
gmm.set_features(features)
log_likelihoods[i] = gmm.train_em()
gmms.append(gmm)
max_idx = log_likelihoods.argmax()
# construct Gaussian mixture components in internal representation
components = []
for i in range(self.num_components):
mu = gmms[max_idx].get_nth_mean(i)
Sigma = gmms[max_idx].get_nth_cov(i)
components.append(Gaussian(mu, Sigma))
# construct a Gaussian mixture model based on the best EM run
pie = gmms[max_idx].get_coef()
proposal = MixtureDistribution(components[0].dimension,
self.num_components, components,
Discrete(pie))
return proposal
def fit_gmm(self, samples):
"""
Runs a couple of em instances on random starting points and returns
internal GMM representation of best instance
"""
features = RealFeatures(samples.T)
gmms = []
log_likelihoods = zeros(self.num_runs_em)
for i in range(self.num_runs_em):
# set up Shogun's GMM class and run em (corresponds to random
# initialisation)
gmm = GMM(self.num_components)
gmm.set_features(features)
log_likelihoods[i] = gmm.train_em()
gmms.append(gmm)
max_idx = log_likelihoods.argmax()
# construct Gaussian mixture components in internal representation
components = []
for i in range(self.num_components):
mu = gmms[max_idx].get_nth_mean(i)
Sigma = gmms[max_idx].get_nth_cov(i)
components.append(Gaussian(mu, Sigma))
# construct a Gaussian mixture model based on the best EM run
pie = gmms[max_idx].get_coef()
proposal = MixtureDistribution(components[0].dimension,
self.num_components, components,
Discrete(pie))
return proposal
def RunGMMShogun():
totalTimer = Timer()
try:
# Load input dataset.
Log.Info("Loading dataset", self.verbose)
dataPoints = np.genfromtxt(self.dataset, delimiter=',')
dataFeat = RealFeatures(dataPoints.T)
# Get all the parameters.
if "gaussians" in options:
g = int(options.pop("gaussians"))
else:
Log.Fatal("Required parameter 'gaussians' not specified!")
raise Exception("missing parameter")
if "max_iterations" in options:
n = int(options.pop("max_iterations"))
else:
n = 0
if len(options) > 0:
Log.Fatal("Unknown parameters: " + str(options))
raise Exception("unknown parameters")
# Create the Gaussian Mixture Model.
model = SGMM(g)
model.set_features(dataFeat)
with totalTimer:
model.train_em(1e-9, n, 1e-9)
except Exception as e:
Log.Info("Exception: " + str(e))
return -1
return totalTimer.ElapsedTime()
def RunGMMShogun(q):
totalTimer = Timer()
try:
# Load input dataset.
Log.Info("Loading dataset", self.verbose)
dataPoints = np.genfromtxt(self.dataset, delimiter=',')
dataFeat = RealFeatures(dataPoints.T)
# Get all the parameters.
g = re.search("-g (\d+)", options)
n = re.search("-n (\d+)", options)
s = re.search("-n (\d+)", options)
g = 1 if not g else int(g.group(1))
n = 250 if not n else int(n.group(1))
# Create the Gaussian Mixture Model.
model = SGMM(g)
model.set_features(dataFeat)
with totalTimer:
model.train_em(1e-9, n, 1e-9)
except Exception as e:
q.put(-1)
return -1
time = totalTimer.ElapsedTime()
q.put(time)
return time
def RunGMMShogun():
totalTimer = Timer()
try:
# Load input dataset.
Log.Info("Loading dataset", self.verbose)
dataPoints = np.genfromtxt(self.dataset, delimiter=',')
dataFeat = RealFeatures(dataPoints.T)
# Get all the parameters.
if "gaussians" in options:
g = int(options.pop("gaussians"))
else:
Log.Fatal("Required parameter 'gaussians' not specified!")
raise Exception("missing parameter")
if "max_iterations" in options:
n = int(options.pop("max_iterations"))
else:
n = 0
if len(options) > 0:
Log.Fatal("Unknown parameters: " + str(options))
raise Exception("unknown parameters")
# Create the Gaussian Mixture Model.
model = SGMM(g)
model.set_features(dataFeat)
with totalTimer:
model.train_em(1e-9, n, 1e-9)
except Exception as e:
Log.Info("Exception: " + str(e))
return -1
return totalTimer.ElapsedTime()
def generate_gmm_classification_data(request):
from modshogun import GMM, Math
num_classes = int(request.POST['num_classes'])
gmm = GMM(num_classes)
total = 40
rng = 4.0
num = total/num_classes
for i in xrange(num_classes):
gmm.set_nth_mean(np.array([Math.random(-rng, rng) for j in xrange(2)]), i)
cov_tmp = Math.normal_random(0.2, 0.1)
cov = np.array([[1.0, cov_tmp], [cov_tmp, 1.0]], dtype=float)
gmm.set_nth_cov(cov, i)
data=[]
labels=[]
for i in xrange(num_classes):
coef = np.zeros(num_classes)
coef[i] = 1.0
gmm.set_coef(coef)
data.append(np.array([gmm.sample() for j in xrange(num)]).T)
labels.append(np.array([i for j in xrange(num)]))
data = np.hstack(data)
data = data / (2.0 * rng)
xmin = np.min(data[0,:])
ymin = np.min(data[1,:])
labels = np.hstack(labels)
toy_data = []
for i in xrange(num_classes*num):
toy_data.append( { 'x': data[0, i] - xmin,
'y': data[1, i] - ymin,
'label': float(labels[i])})
return HttpResponse(json.dumps(toy_data))
def generate_gmm_classification_data(request):
from modshogun import GMM, Math
num_classes = int(request.POST['num_classes'])
gmm = GMM(num_classes)
total = 40
rng = 4.0
num = total / num_classes
for i in xrange(num_classes):
gmm.set_nth_mean(np.array([Math.random(-rng, rng) for j in xrange(2)]),
i)
cov_tmp = Math.normal_random(0.2, 0.1)
cov = np.array([[1.0, cov_tmp], [cov_tmp, 1.0]], dtype=float)
gmm.set_nth_cov(cov, i)
data = []
labels = []
for i in xrange(num_classes):
coef = np.zeros(num_classes)
coef[i] = 1.0
gmm.set_coef(coef)
data.append(np.array([gmm.sample() for j in xrange(num)]).T)
labels.append(np.array([i for j in xrange(num)]))
data = np.hstack(data)
data = data / (2.0 * rng)
xmin = np.min(data[0, :])
ymin = np.min(data[1, :])
labels = np.hstack(labels)
toy_data = []
for i in xrange(num_classes * num):
toy_data.append({
'x': data[0, i] - xmin,
'y': data[1, i] - ymin,
'label': float(labels[i])
})
return HttpResponse(json.dumps(toy_data))
from pylab import figure,show,connect,hist,plot,legend
from numpy import array, append, arange, empty, exp
from modshogun import Gaussian, GMM
from modshogun import RealFeatures
import util
util.set_title('EM for 1d GMM example')
#set the parameters
min_cov=1e-9
max_iter=1000
min_change=1e-9
#setup the real GMM
real_gmm=GMM(3)
real_gmm.set_nth_mean(array([-2.0]), 0)
real_gmm.set_nth_mean(array([0.0]), 1)
real_gmm.set_nth_mean(array([2.0]), 2)
real_gmm.set_nth_cov(array([[0.3]]), 0)
real_gmm.set_nth_cov(array([[0.1]]), 1)
real_gmm.set_nth_cov(array([[0.2]]), 2)
real_gmm.set_coef(array([0.3, 0.5, 0.2]))
#generate training set from real GMM
generated=array([real_gmm.sample()])
for i in range(199):
generated=append(generated, array([real_gmm.sample()]), axis=1)
from modshogun import Gaussian, GMM
from modshogun import RealFeatures
import util
util.set_title('SMEM for 2d GMM example')
#set the parameters
max_iter = 100
max_cand = 5
min_cov = 1e-9
max_em_iter = 1000
min_change = 1e-9
cov_type = 0
#setup the real GMM
real_gmm = GMM(3)
real_gmm.set_nth_mean(array([2.0, 2.0]), 0)
real_gmm.set_nth_mean(array([-2.0, -2.0]), 1)
real_gmm.set_nth_mean(array([2.0, -2.0]), 2)
real_gmm.set_nth_cov(array([[1.0, 0.2], [0.2, 0.5]]), 0)
real_gmm.set_nth_cov(array([[0.2, 0.1], [0.1, 0.5]]), 1)
real_gmm.set_nth_cov(array([[0.3, -0.2], [-0.2, 0.8]]), 2)
real_gmm.set_coef(array([0.3, 0.4, 0.3]))
#generate training set from real GMM
generated = array([real_gmm.sample()])
for i in range(199):
generated = append(generated, array([real_gmm.sample()]), axis=0)
from numpy import array, append, arange, empty, exp
from modshogun import Gaussian, GMM
from modshogun import RealFeatures
import util
util.set_title('SMEM for 1d GMM example')
#set the parameters
max_iter = 100
max_cand = 5
min_cov = 1e-9
max_em_iter = 1000
min_change = 1e-9
#setup the real GMM
real_gmm = GMM(3)
real_gmm.set_nth_mean(array([-2.0]), 0)
real_gmm.set_nth_mean(array([0.0]), 1)
real_gmm.set_nth_mean(array([2.0]), 2)
real_gmm.set_nth_cov(array([[0.3]]), 0)
real_gmm.set_nth_cov(array([[0.1]]), 1)
real_gmm.set_nth_cov(array([[0.2]]), 2)
real_gmm.set_coef(array([0.3, 0.5, 0.2]))
#generate training set from real GMM
generated = array([real_gmm.sample()])
for i in range(199):
generated = append(generated, array([real_gmm.sample()]), axis=1)
from pylab import figure,scatter,contour,show,legend,connect
from numpy import array, append, arange, reshape, empty, exp
from modshogun import Gaussian, GMM
from modshogun import RealFeatures
import util
util.set_title('EM for 2d GMM example')
#set the parameters
min_cov=1e-9
max_iter=1000
min_change=1e-9
cov_type=0
#setup the real GMM
real_gmm=GMM(2)
real_gmm.set_nth_mean(array([1.0, 1.0]), 0)
real_gmm.set_nth_mean(array([-1.0, -1.0]), 1)
real_gmm.set_nth_cov(array([[1.0, 0.2],[0.2, 0.1]]), 0)
real_gmm.set_nth_cov(array([[0.3, 0.1],[0.1, 1.0]]), 1)
real_gmm.set_coef(array([0.3, 0.7]))
#generate training set from real GMM
generated=array([real_gmm.sample()])
for i in range(199):
generated=append(generated, array([real_gmm.sample()]), axis=0)
generated=generated.transpose()
from modshogun import Gaussian, GMM
from modshogun import RealFeatures
import util
util.set_title('SMEM for 2d GMM example')
#set the parameters
max_iter=100
max_cand=5
min_cov=1e-9
max_em_iter=1000
min_change=1e-9
cov_type=0
#setup the real GMM
real_gmm=GMM(3)
real_gmm.set_nth_mean(array([2.0, 2.0]), 0)
real_gmm.set_nth_mean(array([-2.0, -2.0]), 1)
real_gmm.set_nth_mean(array([2.0, -2.0]), 2)
real_gmm.set_nth_cov(array([[1.0, 0.2],[0.2, 0.5]]), 0)
real_gmm.set_nth_cov(array([[0.2, 0.1],[0.1, 0.5]]), 1)
real_gmm.set_nth_cov(array([[0.3, -0.2],[-0.2, 0.8]]), 2)
real_gmm.set_coef(array([0.3, 0.4, 0.3]))
#generate training set from real GMM
generated=array([real_gmm.sample()])
for i in range(199):
generated=append(generated, array([real_gmm.sample()]), axis=0)
from numpy import array, append, arange, empty, exp
from modshogun import Gaussian, GMM
from modshogun import RealFeatures
import util
util.set_title('SMEM for 1d GMM example')
#set the parameters
max_iter=100
max_cand=5
min_cov=1e-9
max_em_iter=1000
min_change=1e-9
#setup the real GMM
real_gmm=GMM(3)
real_gmm.set_nth_mean(array([-2.0]), 0)
real_gmm.set_nth_mean(array([0.0]), 1)
real_gmm.set_nth_mean(array([2.0]), 2)
real_gmm.set_nth_cov(array([[0.3]]), 0)
real_gmm.set_nth_cov(array([[0.1]]), 1)
real_gmm.set_nth_cov(array([[0.2]]), 2)
real_gmm.set_coef(array([0.3, 0.5, 0.2]))
#generate training set from real GMM
generated=array([real_gmm.sample()])
for i in range(199):
generated=append(generated, array([real_gmm.sample()]), axis=1)
from pylab import figure, show, connect, hist, plot, legend
from numpy import array, append, arange, empty, exp
from modshogun import Gaussian, GMM
from modshogun import RealFeatures
import util
util.set_title('EM for 1d GMM example')
#set the parameters
min_cov = 1e-9
max_iter = 1000
min_change = 1e-9
#setup the real GMM
real_gmm = GMM(3)
real_gmm.set_nth_mean(array([-2.0]), 0)
real_gmm.set_nth_mean(array([0.0]), 1)
real_gmm.set_nth_mean(array([2.0]), 2)
real_gmm.set_nth_cov(array([[0.3]]), 0)
real_gmm.set_nth_cov(array([[0.1]]), 1)
real_gmm.set_nth_cov(array([[0.2]]), 2)
real_gmm.set_coef(array([0.3, 0.5, 0.2]))
#generate training set from real GMM
generated = array([real_gmm.sample()])
for i in range(199):
generated = append(generated, array([real_gmm.sample()]), axis=1)
from pylab import figure, scatter, contour, show, legend, connect
from numpy import array, append, arange, reshape, empty, exp
from modshogun import Gaussian, GMM
from modshogun import RealFeatures
import util
util.set_title('EM for 2d GMM example')
#set the parameters
min_cov = 1e-9
max_iter = 1000
min_change = 1e-9
cov_type = 0
#setup the real GMM
real_gmm = GMM(2)
real_gmm.set_nth_mean(array([1.0, 1.0]), 0)
real_gmm.set_nth_mean(array([-1.0, -1.0]), 1)
real_gmm.set_nth_cov(array([[1.0, 0.2], [0.2, 0.1]]), 0)
real_gmm.set_nth_cov(array([[0.3, 0.1], [0.1, 1.0]]), 1)
real_gmm.set_coef(array([0.3, 0.7]))
#generate training set from real GMM
generated = array([real_gmm.sample()])
for i in range(199):
generated = append(generated, array([real_gmm.sample()]), axis=0)
generated = generated.transpose()