def _initModel(self, cacheRoot):
"""
Initialize the network; self.networdDataPath must already be set.
"""
encoder = CioEncoder(retinaScaling=self.retinaScaling,
retina=self.retina,
apiKey=self.apiKey,
maxSparsity=self.maxSparsity,
verbosity=self.verbosity-1,
cacheDir=cacheRoot)
# This encoder specifies the LanguageSensor output width.
return createAndConfigureNetwork(None, self.networkConfig, encoder)
def _initModel(self, cacheRoot):
"""
Initialize the network; self.networdDataPath must already be set.
"""
encoder = CioEncoder(retinaScaling=self.retinaScaling,
retina=self.retina,
apiKey=self.apiKey,
maxSparsity=self.maxSparsity,
verbosity=self.verbosity - 1,
cacheDir=cacheRoot)
# This encoder specifies the LanguageSensor output width.
return createAndConfigureNetwork(None, self.networkConfig, encoder)
def _initModel(self, k):
"""
Initialize the network
"""
encoder = CioEncoder(retinaScaling=self.retinaScaling,
retina=self.retina,
fingerprintType=EncoderTypes.document,
apiKey=self.apiKey,
verbosity=self.verbosity-1)
modelConfig["classifierRegionConfig"]["regionParams"]["k"] = k
modelConfig["classifierRegionConfig"]["regionParams"][
"maxCategoryCount"] = self.numLabels
self.networkConfig = modelConfig
self.network = createAndConfigureNetwork(None, self.networkConfig, encoder)
def _initModel(self, k):
"""
Initialize the network
"""
encoder = CioEncoder(retinaScaling=self.retinaScaling,
retina=self.retina,
fingerprintType=EncoderTypes.document,
apiKey=self.apiKey,
verbosity=self.verbosity - 1)
modelConfig["classifierRegionConfig"]["regionParams"]["k"] = k
modelConfig["classifierRegionConfig"]["regionParams"][
"maxCategoryCount"] = self.numLabels
self.networkConfig = modelConfig
self.network = createAndConfigureNetwork(None, self.networkConfig,
encoder)
def runNetwork(networkConfig, filePath, runClustering):
dataSource = FileRecordStream(streamID=filePath)
network = createAndConfigureNetwork(dataSource, networkConfig)
(sensorRegion,
spRegion,
tmRegion,
tpRegion,
classifierRegion) = setRegionLearning(network, networkConfig)
trace = initTrace(runClustering)
numCells = networkConfig['tmRegionConfig']['regionParams']['inputWidth'] * \
networkConfig['tmRegionConfig']['regionParams']['cellsPerColumn']
if runClustering:
clustering = Clustering(numCells,
MERGE_THRESHOLD,
ANOMALOUS_THRESHOLD,
STABLE_THRESHOLD,
MIN_CLUSTER_SIZE,
SIMILARITY_THRESHOLD)
recordNumber = 0
while 1:
try:
network.run(1)
(sensorValue,
actualCategory,
tmActiveCells,
tmPredictedActiveCells,
rawAnomalyScore,
rollingAnomalyScore,
tpActiveCells,
classificationInference,
rawClassificationAccuracy,
rollingClassificationAccuracy) = computeNetworkStats(trace,
ROLLING_ACCURACY_WINDOW,
sensorRegion,
tmRegion,
tpRegion,
classifierRegion)
trace = updateTrace(trace,
recordNumber,
sensorValue,
actualCategory,
tmActiveCells,
tmPredictedActiveCells,
rawAnomalyScore,
rollingAnomalyScore,
tpActiveCells,
classificationInference,
rawClassificationAccuracy,
rollingClassificationAccuracy)
if recordNumber % 500 == 0:
outputClassificationInfo(recordNumber,
sensorValue,
actualCategory,
rollingAnomalyScore,
classificationInference,
rollingClassificationAccuracy)
if runClustering:
if CELLS_TO_CLUSTER == 'tmActiveCells':
tmCells = tmActiveCells
elif CELLS_TO_CLUSTER == 'tmPredictedActiveCells':
tmCells = tmPredictedActiveCells
else:
raise ValueError(
'CELLS_TO_CLUSTER value can only be "tmActiveCells" '
'or "tmPredictedActiveCells" but is %s'
% CELLS_TO_CLUSTER)
if ANOMALY_SCORE == 'rollingAnomalyScore':
anomalyScore = rollingAnomalyScore
elif ANOMALY_SCORE == 'rawAnomalyScore':
anomalyScore = rawAnomalyScore
else:
raise ValueError('ANOMALY_SCORE value can only be "rawAnomalyScore" '
'or "rollingAnomalyScore" but is %s'
% ANOMALY_SCORE)
(predictedCluster,
clusteringConfidence) = clustering.cluster(recordNumber,
tmCells,
anomalyScore,
actualCategory)
(clusteringInference,
predictedClusterId,
rawClusteringAccuracy,
rollingClusteringAccuracy,
clusterHomogeneity) = computeClusteringStats(trace,
predictedCluster,
clustering,
actualCategory)
trace = updateClusteringTrace(trace,
clusteringInference,
predictedClusterId,
rawClusteringAccuracy,
rollingClusteringAccuracy,
clusterHomogeneity,
clusteringConfidence)
if recordNumber % 100 == 0:
numClusters = len(clustering.getClusters())
outputClusteringInfo(clusteringInference,
predictedClusterId,
rollingClusteringAccuracy,
clusterHomogeneity,
clusteringConfidence,
numClusters)
recordNumber += 1
except StopIteration:
print "Data streaming completed!"
break
if runClustering:
outputClustersStructure(clustering)
outputInterClusterDist(clustering, numCells)
return trace, recordNumber
def testClassificationAccuracy(self):
"""Test classification accuracy for sensor data."""
networkConfigurations = generateSampleNetworkConfig(
self.templateNetworkConfig, NUM_CATEGORIES)
for networkConfig in networkConfigurations:
for noiseAmplitude in WHITE_NOISE_AMPLITUDES:
for signalMean in SIGNAL_MEANS:
for signalAmplitude in SIGNAL_AMPLITUDES:
for signalPeriod in SIGNAL_PERIODS:
sensorType = networkConfig[
"sensorRegionConfig"].get("regionType")
spEnabled = networkConfig[
"sensorRegionConfig"].get("regionEnabled")
tmEnabled = networkConfig["tmRegionConfig"].get(
"regionEnabled")
upEnabled = networkConfig["tpRegionConfig"].get(
"regionEnabled")
classifierType = networkConfig[
"classifierRegionConfig"].get("regionType")
expParams = ("RUNNING EXPERIMENT WITH PARAMS:\n"
" * numRecords=%s\n"
" * signalAmplitude=%s\n"
" * signalMean=%s\n"
" * signalPeriod=%s\n"
" * noiseAmplitude=%s\n"
" * sensorType=%s\n"
" * spEnabled=%s\n"
" * tmEnabled=%s\n"
" * tpEnabled=%s\n"
" * classifierType=%s\n") % (
NUM_RECORDS, signalAmplitude,
signalMean, signalPeriod,
noiseAmplitude,
sensorType.split(".")[1],
spEnabled, tmEnabled, upEnabled,
classifierType.split(".")[1])
print expParams
inputFile = generateSensorData(
DATA_DIR, OUTFILE_NAME, signalMean,
signalPeriod, SEQUENCE_LENGTH, NUM_RECORDS,
signalAmplitude, NUM_CATEGORIES,
noiseAmplitude)
dataSource = FileRecordStream(streamID=inputFile)
network = createAndConfigureNetwork(
dataSource, networkConfig)
partitions = generateNetworkPartitions(
networkConfig, NUM_RECORDS)
classificationAccuracy = trainNetwork(
network, networkConfig, partitions,
NUM_RECORDS)
if (noiseAmplitude == 0 and signalMean == 1.0
and signalAmplitude == 1.0
and signalPeriod == 20.0
and classifierType == KNN_CLASSIFIER_TYPE
and spEnabled and tmEnabled
and not upEnabled):
self.assertEqual(classificationAccuracy,
100.00)
elif (noiseAmplitude == 0 and signalMean == 1.0
and signalAmplitude == 1.0
and signalPeriod == 20.0
and classifierType == CLA_CLASSIFIER_TYPE
and spEnabled and tmEnabled
and not upEnabled):
self.assertEqual(classificationAccuracy,
100.00)
elif (noiseAmplitude == 0 and signalMean == 1.0
and signalAmplitude == 1.0
and signalPeriod == 20.0
and classifierType == CLA_CLASSIFIER_TYPE
and spEnabled and not tmEnabled
and not upEnabled):
self.assertEqual(classificationAccuracy,
100.00)
elif (noiseAmplitude == 1.0 and signalMean == 1.0
and signalAmplitude == 1.0
and signalPeriod == 20.0
and classifierType == CLA_CLASSIFIER_TYPE
and spEnabled and tmEnabled
and not upEnabled):
# using AlmostEqual until the random bug issue is fixed
self.assertAlmostEqual(classificationAccuracy,
80,
delta=5)
elif (noiseAmplitude == 1.0 and signalMean == 1.0
and signalAmplitude == 1.0
and signalPeriod == 20.0
and classifierType == CLA_CLASSIFIER_TYPE
and spEnabled and not tmEnabled
and not upEnabled):
# using AlmostEqual until the random bug issue is fixed
self.assertAlmostEqual(classificationAccuracy,
81,
delta=5)
def testClassificationAccuracy(self):
"""Test classification accuracy for sensor data."""
networkConfigurations = generateSampleNetworkConfig(
self.templateNetworkConfig, NUM_CATEGORIES)
for networkConfig in networkConfigurations:
for noiseAmplitude in WHITE_NOISE_AMPLITUDES:
for signalMean in SIGNAL_MEANS:
for signalAmplitude in SIGNAL_AMPLITUDES:
for signalPeriod in SIGNAL_PERIODS:
sensorType = networkConfig["sensorRegionConfig"].get(
"regionType")
spEnabled = networkConfig["sensorRegionConfig"].get(
"regionEnabled")
tmEnabled = networkConfig["tmRegionConfig"].get(
"regionEnabled")
upEnabled = networkConfig["tpRegionConfig"].get(
"regionEnabled")
classifierType = networkConfig["classifierRegionConfig"].get(
"regionType")
expParams = ("RUNNING EXPERIMENT WITH PARAMS:\n"
" * numRecords=%s\n"
" * signalAmplitude=%s\n"
" * signalMean=%s\n"
" * signalPeriod=%s\n"
" * noiseAmplitude=%s\n"
" * sensorType=%s\n"
" * spEnabled=%s\n"
" * tmEnabled=%s\n"
" * tpEnabled=%s\n"
" * classifierType=%s\n"
) % (NUM_RECORDS,
signalAmplitude,
signalMean,
signalPeriod,
noiseAmplitude,
sensorType.split(".")[1],
spEnabled,
tmEnabled,
upEnabled,
classifierType.split(".")[1])
print expParams
inputFile = generateSensorData(DATA_DIR,
OUTFILE_NAME,
signalMean,
signalPeriod,
SEQUENCE_LENGTH,
NUM_RECORDS,
signalAmplitude,
NUM_CATEGORIES,
noiseAmplitude)
dataSource = FileRecordStream(streamID=inputFile)
network = createAndConfigureNetwork(dataSource,
networkConfig)
partitions = generateNetworkPartitions(networkConfig,
NUM_RECORDS)
classificationAccuracy = trainNetwork(network, networkConfig,
partitions, NUM_RECORDS)
if (noiseAmplitude == 0
and signalMean == 1.0
and signalAmplitude == 1.0
and signalPeriod == 20.0
and classifierType == KNN_CLASSIFIER_TYPE
and spEnabled
and tmEnabled
and not upEnabled):
self.assertEqual(classificationAccuracy, 100.00)
elif (noiseAmplitude == 0
and signalMean == 1.0
and signalAmplitude == 1.0
and signalPeriod == 20.0
and classifierType == CLA_CLASSIFIER_TYPE
and spEnabled
and tmEnabled
and not upEnabled):
self.assertEqual(classificationAccuracy, 100.00)
elif (noiseAmplitude == 0
and signalMean == 1.0
and signalAmplitude == 1.0
and signalPeriod == 20.0
and classifierType == CLA_CLASSIFIER_TYPE
and spEnabled
and not tmEnabled
and not upEnabled):
self.assertEqual(classificationAccuracy, 100.00)
elif (noiseAmplitude == 1.0
and signalMean == 1.0
and signalAmplitude == 1.0
and signalPeriod == 20.0
and classifierType == CLA_CLASSIFIER_TYPE
and spEnabled
and tmEnabled
and not upEnabled):
# using AlmostEqual until the random bug issue is fixed
self.assertAlmostEqual(classificationAccuracy, 80, delta=5)
elif (noiseAmplitude == 1.0
and signalMean == 1.0
and signalAmplitude == 1.0
and signalPeriod == 20.0
and classifierType == CLA_CLASSIFIER_TYPE
and spEnabled
and not tmEnabled
and not upEnabled):
# using AlmostEqual until the random bug issue is fixed
self.assertAlmostEqual(classificationAccuracy, 81, delta=5)
def createNetwork(dataSource, networkConfig, serializedModelPath):
if serializedModelPath:
return loadNetwork(serializedModelPath, dataSource)
else:
return createAndConfigureNetwork(dataSource, networkConfig)
def createNetwork(dataSource, networkConfig, serializedModelPath):
if serializedModelPath:
return loadNetwork(serializedModelPath, dataSource)
else:
return createAndConfigureNetwork(dataSource, networkConfig)