pip install nupic
from nupic.encoders import RandomDistributedScalarEncoder
from nupic.algorithms import anomaly_likelihood
from nupic.algorithms import htm_prediction
from nupic.algorithms import spatial_pooler
from nupic.algorithms import temporal_memory
from nupic.algorithms import anomaly
from nupic.data.plotter import Plotter
inputData = [1, 2, 3, 4, 5, 6, 5, 4, 3, 2]
sp = spatial_pooler.SpatialPooler()
enc = RandomDistributedScalarEncoder()
spatialInput = enc.encode(inputData)
sp.compute(spatialInput, learn=True)
tm = temporal_memory.TemporalMemory()
anomaly_likelihood = anomaly_likelihood.AnomalyLikelihood()
for step in range(len(inputData)):
spatialInput = enc.encode(inputData[step])
predictedActiveCells = htm_prediction.Prediction.getMostProbableCells()
predictive = len(predictedActiveCells) > 0
tm.compute(predictive, activeColumns)
anomaly_score = anomaly_likelihood.anomaly
anomaly_likelihood.anomaly = anomaly_score
if anomaly_score > 0.5:
print("Anomaly detected at step", step)
predictedValue = tm.infer()
print("Predicted value at step", step, ":", predictedValue)
steps = list(range(len(inputData)))
Plotter.plot(data={"Input Data": inputData}, steps=steps)
Plotter.plot(data={"Predicted Value": predictedValues}, steps=steps)
