Python KNNClassifierRegion.KNNClassifierRegion類代碼示例

  def __init__(self,
               trainRecords,
               anomalyThreshold,
               cacheSize,
               classificationVectorType=1,
               activeColumnCount=40,
               classificationMaxDist=0.30,
               **classifierArgs):

    # Internal Region Values
    self._maxLabelOutputs = 16
    self._activeColumnCount = activeColumnCount
    self._prevPredictedColumns = numpy.array([])
    self._anomalyVectorLength = None
    self._classificationMaxDist = classificationMaxDist
    self._iteration = 0

    # Set to create deterministic classifier
    classifierArgs['SVDDimCount'] = None

    # Parameters
    self.trainRecords = trainRecords
    self.anomalyThreshold = anomalyThreshold
    self.cacheSize = cacheSize
    self.classificationVectorType = classificationVectorType

    self._knnclassifierArgs = classifierArgs
    self._knnclassifier = KNNClassifierRegion(**self._knnclassifierArgs)
    self.labelResults = []
    self.saved_categories = []
    self._recordsCache = []

    self._version = KNNAnomalyClassifierRegion.__VERSION__

  def __setstate__(self, state):
    """
    Set the state of ourself from a serialized state.
    """
    if '_version' not in state or state['_version'] == 1:

      knnclassifierProps = state.pop('_knnclassifierProps')

      self.__dict__.update(state)
      self._knnclassifier = KNNClassifierRegion(**self._knnclassifierArgs)
      self._knnclassifier.__setstate__(knnclassifierProps)

      self._version = KNNAnomalyClassifierRegion.__VERSION__
    else:
      raise Exception("Invalid KNNAnomalyClassifierRegion version. Current "
          "version: %s" % (KNNAnomalyClassifierRegion.__VERSION__))

  def __setstate__(self, state):
    """
    Set the state of ourself from a serialized state.
    """
    if '_version' not in state or state['_version'] == 1:

      knnclassifierProps = state.pop('_knnclassifierProps')

      self.__dict__.update(state)
      self._knnclassifier = KNNClassifierRegion(**self._knnclassifierArgs)
      self._knnclassifier.__setstate__(knnclassifierProps)

      self._version = KNNAnomalyClassifierRegion.__VERSION__
    else:
      raise Exception("Invalid KNNAnomalyClassifierRegion version. Current "
          "version: %s" % (KNNAnomalyClassifierRegion.__VERSION__))


    # --------------------------------------------------------------------------
    # Migrate from when Anomaly was separated to stand-alone class
    if not hasattr(self, "_anomaly"):
      self._anomaly = Anomaly()

#.........這裏部分代碼省略.........
            isDefaultInput=True,
            requireSplitterMap=False),

          tpLrnActiveStateT=dict(
            description="""Active cells in the learn state at time T from TP.
                        This is used to classify on.""",
            dataType='Real32',
            count=0,
            required=True,
            regionLevel=False,
            isDefaultInput=True,
            requireSplitterMap=False)
        ),

        outputs=dict(
        ),

        parameters=dict(
          trainRecords=dict(
            description='Number of records to wait for training',
            dataType='UInt32',
            count=1,
            constraints='',
            defaultValue=0,
            accessMode='Create'),

          anomalyThreshold=dict(
            description='Threshold used to classify anomalies.',
            dataType='Real32',
            count=1,
            constraints='',
            defaultValue=0,
            accessMode='Create'),

          cacheSize=dict(
            description='Number of records to store in cache.',
            dataType='UInt32',
            count=1,
            constraints='',
            defaultValue=0,
            accessMode='Create'),

          classificationVectorType=dict(
            description="""Vector type to use when classifying.
              1 - Vector Column with Difference (TP and SP)
            """,
            dataType='UInt32',
            count=1,
            constraints='',
            defaultValue=1,
            accessMode='ReadWrite'),

          activeColumnCount=dict(
            description="""Number of active columns in a given step. Typically
            equivalent to SP.numActivePerInhArea""",
            dataType='UInt32',
            count=1,
            constraints='',
            defaultValue=40,
            accessMode='ReadWrite'),

          classificationMaxDist=dict(
            description="""Maximum distance a sample can be from an anomaly
            in the classifier to be labeled as an anomaly.

            Ex: With rawOverlap distance, a value of 0.65 means that the points
            must be at most a distance 0.65 apart from each other. This
            translates to they must be at least 35% similar.""",
            dataType='Real32',
            count=1,
            constraints='',
            defaultValue=0.65,
            accessMode='Create'
            )
        ),
        commands=dict(
          getLabels=dict(description=
            "Returns a list of label dicts with properties ROWID and labels."
            "ROWID corresponds to the records id and labels is a list of "
            "strings representing the records labels.  Takes additional "
            "integer properties start and end representing the range that "
            "will be returned."),

          addLabel=dict(description=
            "Takes parameters start, end and labelName. Adds the label "
            "labelName to the records from start to end. This will recalculate "
            "labels from end to the most recent record."),

          removeLabels=dict(description=
            "Takes additional parameters start, end, labelFilter.  Start and "
            "end correspond to range to remove the label. Remove labels from "
            "each record with record ROWID in range from start to end, "
            "noninclusive of end. Removes all records if labelFilter is None, "
            "otherwise only removes the labels eqaul to labelFilter.")
        )
      )

    ns['parameters'].update(KNNClassifierRegion.getSpec()['parameters'])

    return ns

class KNNAnomalyClassifierRegion(PyRegion):
  """
  KNNAnomalyClassifierRegion wraps the KNNClassifierRegion to classify clamodel
  state.  It allows for individual records to be classified as anomalies and
  supports anomaly detection even after the model has learned the anomalous
  sequence.

  Methods:
    compute() - called by clamodel during record processing
    getLabels() - return points with classification records
    addLabel() - add a set label to a given set of points
    removeLabels() - remove labels from a given set of points

  Parameters:
    trainRecords - number of records to skip before classification
    anomalyThreshold - threshold on anomaly score to automatically classify
                       record as an anomaly
    cacheSize - number of records to keep in cache. Can only recalculate
                records kept in cache when setting the trainRecords.

  """

  @classmethod
  def getSpec(cls):
    ns = dict(
        description=KNNAnomalyClassifierRegion.__doc__,
        singleNodeOnly=True,
        inputs=dict(
          spBottomUpOut=dict(
            description="""The output signal generated from the bottom-up inputs
                            from lower levels.""",
            dataType='Real32',
            count=0,
            required=True,
            regionLevel=False,
            isDefaultInput=True,
            requireSplitterMap=False),

          tpTopDownOut=dict(
            description="""The top-down inputsignal, generated from
                          feedback from upper levels""",
            dataType='Real32',
            count=0,
            required=True,
            regionLevel=False,
            isDefaultInput=True,
            requireSplitterMap=False),

          tpLrnActiveStateT=dict(
            description="""Active cells in the learn state at time T from TP.
                        This is used to classify on.""",
            dataType='Real32',
            count=0,
            required=True,
            regionLevel=False,
            isDefaultInput=True,
            requireSplitterMap=False)
        ),

        outputs=dict(
        ),

        parameters=dict(
          trainRecords=dict(
            description='Number of records to wait for training',
            dataType='UInt32',
            count=1,
            constraints='',
            defaultValue=0,
            accessMode='Create'),

          anomalyThreshold=dict(
            description='Threshold used to classify anomalies.',
            dataType='Real32',
            count=1,
            constraints='',
            defaultValue=0,
            accessMode='Create'),

          cacheSize=dict(
            description='Number of records to store in cache.',
            dataType='UInt32',
            count=1,
            constraints='',
            defaultValue=0,
            accessMode='Create'),

          classificationVectorType=dict(
            description="""Vector type to use when classifying.
              1 - Vector Column with Difference (TP and SP)
            """,
            dataType='UInt32',
            count=1,
            constraints='',
            defaultValue=1,
            accessMode='ReadWrite'),

          activeColumnCount=dict(
            description="""Number of active columns in a given step. Typically
            equivalent to SP.numActivePerInhArea""",
#.........這裏部分代碼省略.........