def _initializeGoodModels(data, numModels, states, obs):
"""Initialization technique for selecting a good pool of models by which
to start the expectation maximization algorithm for unsupervised learning
"""
sigma = IntegerRange(0, obs)
models = []
distances = numpy.zeros((len(data)), float)
# Select initial random sequence
seq = data[int(random.random() * len(data))]
models.append(hmmextra.obsToModel(seq, states, obs))
for i in range(numModels - 1):
for j in range(len(data)):
tmp = hmmextra.obsToModel(data[j], states, obs)
tmpDistances = hmmextra.hmmDistAll(data[j], models, sigma)
tmpDistances.sort()
distances[j] = tmpDistances[0]
distances /= sum(distances)
# Select a model with probability dist/sum(all distance)
val = random.random()
distSum = 0
for j in range(len(distances)):
distSum += distances[j]
if distSum >= val:
# Make a new model using this data element
models.append(hmmextra.obsToModel(data[j], states, obs))
break
return models
def _initializeGoodModels(data, numModels, states, obs):
"""Initialization technique for selecting a good pool of models by which
to start the expectation maximization algorithm for unsupervised learning
"""
sigma = IntegerRange(0, obs)
models = []
distances = numpy.zeros((len(data)), float)
#Select initial random sequence
seq = data[int(random.random() * len(data))]
models.append(hmmextra.obsToModel(seq, states, obs))
for i in range(numModels - 1):
for j in range(len(data)):
tmp = hmmextra.obsToModel(data[j], states, obs)
tmpDistances = hmmextra.hmmDistAll(data[j], models, sigma)
tmpDistances.sort()
distances[j] = tmpDistances[0]
distances /= sum(distances)
#Select a model with probability dist/sum(all distance)
val = random.random()
distSum = 0
for j in range(len(distances)):
distSum += distances[j]
if distSum >= val:
#Make a new model using this data element
models.append(hmmextra.obsToModel(data[j], states, obs))
break
return models
def drawHMMCluster(data, models, numSensors, \
writeLocation = "../output/out.png", \
spacing = 10, numberBase = 2, \
scaling = 5, \
cColor = (0, 255, 0), cCluster = (0, 0, 0)):
"""Draw all data associated with a given hidden markov model.
Space the data by the value spacing.
"""
lenInstances = 0
#Get length of each image.
for d in data:
lenInstances += len(d)
#Make the image
iLen = lenInstances * scaling + spacing * len(data)
if iLen == 0:
iLen = 1
im = Image.new("RGB", ((numSensors + 1) * scaling + 50, iLen))
d = ImageDraw.Draw(im)
currentY = 0
#Draw the instances
for i in range(len(data)):
#Invert the data to an array.
tmpData = bbdata.uncompressData(data[i], numSensors, numberBase)
#Caluclate values
sigma = IntegerRange(0, numberBase**numSensors)
values = hmmextra.hmmDistAll(data[i], models, sigma)
values.sort()
_drawArrayText(d,
tmpData,
0,
currentY,
values,
scale=scaling,
cCluster=cCluster)
currentY += tmpData.shape[0] * scaling + spacing
im.save(writeLocation, "PNG")
def drawHMMCluster(data, models, numSensors, \
writeLocation = "../output/out.png", \
spacing = 10, numberBase = 2, \
scaling = 5, \
cColor = (0, 255, 0), cCluster = (0, 0, 0)):
"""Draw all data associated with a given hidden markov model.
Space the data by the value spacing.
"""
lenInstances = 0
#Get length of each image.
for d in data:
lenInstances += len(d)
#Make the image
iLen = lenInstances * scaling + spacing * len(data)
if iLen == 0:
iLen = 1
im = Image.new("RGB", ((numSensors + 1) * scaling + 50, iLen))
d = ImageDraw.Draw(im)
currentY = 0
#Draw the instances
for i in range(len(data)):
#Invert the data to an array.
tmpData = bbdata.uncompressData(data[i], numSensors, numberBase)
#Caluclate values
sigma = IntegerRange(0, numberBase**numSensors)
values = hmmextra.hmmDistAll(data[i], models, sigma)
values.sort()
_drawArrayText(d, tmpData, 0, currentY, values, scale = scaling, cCluster=cCluster)
currentY += tmpData.shape[0] * scaling + spacing
im.save(writeLocation, "PNG")