Sounds interesting. Wont that depend on the data the model was trained with and the model settings [currently provided by the swarm process]. I assume the benchmarking tool will be a interface that all Numenta compliant implementations will have to implement.
I like the concept Chandan On Tue, Jan 6, 2015 at 8:33 AM, Matthew Taylor <[email protected]> wrote: > On Tue, Jan 6, 2015 at 7:53 AM, Matt Keith <[email protected]> wrote: > > That makes me think, do we have a standard benchmark defined to test an > HTM > > implementation against? > > We are building an anomaly detection benchmark at > https://github.com/numenta/NAB. Other anomaly detection algorithms can > be plugged into this framework. > > --------- > Matt Taylor > OS Community Flag-Bearer > Numenta > >
