niml.model.nisclassifier package

Submodules

niml.model.nisclassifier.classifier module

class niml.model.nisclassifier.classifier.Classifier(cls_state=None, sdr_width=None, subclass_thresh=None, unknown_thresh=None, min_overlap=None)

Processes oSDR data to create classification signatures during fit, and performs predictions during evaluate and predict. The evaluate method will return a dictionary of scikit-learn metrics while predict returns a list of classifications for each oSDR

Parameters

  • sdr_width (integer) – the largest symbol an SDR might contain+1, typically the number of neurons in the pooler.

  • subclass_thresh (float) – A value that directs the model when to create a new subclass during fit. Anything below the threshold will cause the creation of a new subclass

  • unknown_thresh (float) – A value that indicates when the model needs to create a new subclass during continuous learning. Anything below the threshold will cause the creation of a new subclass (only applies when learning is on)

  • min_overlap (float) – A value that represents the minimum percentage of overlap between an output SDR and subclass signature to determine if the item represented by the output SDR should be considered a member of the subclass

fit(data, labels)

Train the classifier using the provided data and labels

Parameters

  • data (list) – oSDR data generated from a pooler process

  • labels (list) – Observation ground-truth labels

Returns

  • True if a valid classifier exists and we were able to process the

  • training data. False otherwise.

predict(sdrs)

Generates a list of top-level classification labels for the oSDR sets passed in by passing them to predict_subclass and slicing off the label from each tuple provided from that function.

Parameters

sdrs (list) – oSDRs against which to make predictions

Return type

List (possibly empty) of predicted labels for each SDR

save_classifier(output_file=None)

Provides a means for saving the current state of the classifier

Parameters

output_file (string) – file name string

Returns

  • True if a valid classifier exists and we were able to save the

  • configuration to the specified file. False otherwise.