There is also an implementation in https://github.com/numenta/nupic.rogue for a general purpose metric collection agent that can be used with nupic, grok, or htmengine for the processing.
You could add additional metrics in https://github.com/numenta/nupic.rogue/blob/master/avogadro/network_agent.py to look for more specific metrics. On Thu, Aug 20, 2015 at 10:16 AM, Matthew Taylor <[email protected]> wrote: > Richard, > > This is entirely possible with HTM & NuPIC. Have you seen our tutorials > [1]? I suggest the Hot Gym Anomaly Tutorial. If you want to run lots of > models, have a look at the HTM Engine tutorial for traffic anomalies. > > [1] https://github.com/numenta/nupic/wiki/Using-NuPIC#tutorials > > > --------- > Matt Taylor > OS Community Flag-Bearer > Numenta > > On Thu, Aug 20, 2015 at 12:31 AM, Pascal Weinberger < > [email protected]> wrote: > >> Hey, >> >> Welcome! >> >> Do you know about Grok? :) >> I pretty much does the job, so Jap it's totally possible :) >> http://numenta.com/grok/ >> On Aug 20, 2015 5:07 AM, "Richard Compton" <[email protected]> >> wrote: >> >>> Hi, >>> I am interested in using nupic to detect anomalies in network traffic >>> specifically to identify potential security issues. I use an open source >>> application called bro that continuously outputs a log file that has >>> timestamp, source IP/port and destination IP/port, bytes transferred and >>> some other data for every network connection it sees. >>> I am completely new to nupic and machine learning. Is it possible to >>> feed nupic this information and have it identify when there is anomalous >>> network traffic that does not follow the normal patterns of traffic for >>> that time period? I do have some experience programming with python. >>> Thanks in advance! >>> >>> -- >>> Thanks, >>> Rich Compton >>> >> >
