Usability

Newly Updated! (02/22/2015)

See new usage example of complete module assembly below!

#The major components comprising the full usability stack of NuPIC components is now complete and available!

#See the blog post here

Normal Usage

Combinations of the following can be assemble together in your code, to yield quality predictive results these are:

• Encoders [see the full list of available encoders here] (https://github.com/numenta/htm.java/tree/master/src/main/java/org/numenta/nupic/encoders)

• SpatialPooler here

• TemporalMemory here

• Classifier here

• Anomaly (detection and prediction) here ... and for some more information see the blog

The stack of typical components consists of the above 4 or 5, but one or more of them can be eliminated or added to in order to achieve the results you require.

In addition, an example of SpatialPooler usage can be found in this example

The normal pattern of usage is to feed your data into a configured Encoder, which would then produce an SDR (Sparse Data Representation). More information on SDR's can be found here

Following this, the SDR is then fed into a SpatialPooler, and the output of the SpatialPooler is fed into the TemporalPooler which then outputs SDR's representing predicted input of the next inferencing cycle. In order to relate this "prediction" to a result in the context of the input, in some cases a Classifier is used. In addition, the Anomaly computation class can be used for both anomaly detection and anomaly prediction.

New usage example code!

For an example of tying together components and overall usage, see this Quick & Dirty example.

This wiki is currently under construction as the pace of HTM.java's growth and development has rocketed! If you have a knack for crafty documentation, HTM.java can use your help! Please feel free to contact us

For more information about upcoming development, see: the Roadmap

Next: Architecture