Re: Using anomalyProbability gives TypeError: cannot perform reduce with flexible type

[Wakan Tanka]

Thank you Matt,
this was the case. Two questions:

1. I am using both shifted and unshifted inferences (col3 and col4). 
Should I pass anomalyScore from (un)shifted to anomalyProbability?

2. Is there some documents regarding anomaly_likelihood and 
anomalyProbability()?

Thank you


On 02/15/2016 05:23 PM, Matthew Taylor wrote:
At first glance, you might simply need to cast "col2" into a float:
float(col2). It looks like it might be a string?
---------
Matt Taylor
OS Community Flag-Bearer
Numenta


On Sun, Feb 14, 2016 at 5:36 AM, Wakan Tanka <wakat...@gmail.com> wrote:
Here is full trace:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
/home/wakatana/experiments_today/v3/run_nupic.py in <module>()
     239     SWARM_CFG["PREDICTION_STEP"],                      # PREDICTION
STEP
     240     Verbose=True,                      # VERBOSE
--> 241     VeryVerbose=False                     # VERY VERBOSE
     242   )
     243   RUNMODEL_STOP_TIME = SCRIPT_STOP_TIME =
calendar.timegm(time.gmtime())

/home/wakatana/experiments_today/v3/experiments/hot_gym_anomaly/run_model/run_model.py
in runModel(model, inputFile, outputFile, predictionSteps, Verbose,
VeryVerbose)
      67
      68     # Compute the Anomaly Likelihood
---> 69     likelihood    = anomalyLikelihood.anomalyProbability(col2, tmp,
col1)
      70     logLikelihood =
anomalyLikelihood.computeLogLikelihood(likelihood)
      71

/home/wakatana/.local/lib/python2.7/site-packages/nupic-0.3.0.dev0-py2.7-linux-x86_64.egg/nupic/algorithms/anomaly_likelihood.pyc
in anomalyProbability(self, value, anomalyScore, timestamp)
     140           estimateAnomalyLikelihoods(
     141             self._historicalScores,
--> 142             skipRecords = self._claLearningPeriod)
     143           )
     144

/home/wakatana/.local/lib/python2.7/site-packages/nupic-0.3.0.dev0-py2.7-linux-x86_64.egg/nupic/algorithms/anomaly_likelihood.pyc
in estimateAnomalyLikelihoods(anomalyScores, averagingWindow, skipRecords,
verbosity)
     297     metricValues = numpy.array(s)
     298     metricDistribution = estimateNormal(metricValues[skipRecords:],
--> 299 performLowerBoundCheck=False)
     300
     301     if metricDistribution["variance"] < 1.5e-5:

/home/wakatana/.local/lib/python2.7/site-packages/nupic-0.3.0.dev0-py2.7-linux-x86_64.egg/nupic/algorithms/anomaly_likelihood.pyc
in estimateNormal(sampleData, performLowerBoundCheck)
     511   params = {
     512       "name": "normal",
--> 513       "mean": numpy.mean(sampleData),
     514       "variance": numpy.var(sampleData),
     515   }

/usr/lib/python2.7/dist-packages/numpy/core/fromnumeric.pyc in mean(a, axis,
dtype, out, keepdims)
    2714
    2715     return _methods._mean(a, axis=axis, dtype=dtype,
-> 2716                             out=out, keepdims=keepdims)
    2717
    2718 def std(a, axis=None, dtype=None, out=None, ddof=0, keepdims=False):

/usr/lib/python2.7/dist-packages/numpy/core/_methods.pyc in _mean(a, axis,
dtype, out, keepdims)
      60         dtype = mu.dtype('f8')
      61
---> 62     ret = um.add.reduce(arr, axis=axis, dtype=dtype, out=out,
keepdims=keepdims)
      63     if isinstance(ret, mu.ndarray):
      64         ret = um.true_divide(


TypeError: cannot perform reduce with flexible type