I have read and thought about this a bit. Here is my version, please
comment.

*Level 1:* Single device level analytics. Show the data in graph, map, send
an Alarm it is too high .. etc ( we have this mostly)

Tasks

   1. Inbuilt analytics for well known properties like location etc
   2. Integrate Gadget generation wizard so users can build their own
   graphs with support for realtime graphs
   3. Setup alerts from the charts or UI ( this is realtime analytics)


*Level 2:* analytics across multiple devices - all, groups, hierarchy ..
aggregate across

   1. Add support for groups and aggregation for Gadget generation wizard
   integration.
   2. Support to write custom queries if needed


*Level 3: *Predictive analytics - this is three things, a) classify b)
predict next value or  c) detect anomalies. We should be able to let them
do via wizard .. at least basic version.

Ability to select a set of data and jump to ML Wizard for exploration and
processing


   1. Classification - apply available algorithm, build a model, and either
   publish to registry or setup query/ alert using the model
   2. Next value - same as above. Need time series support as well.
   3. Anomaly detection -  integrate with Seshika’s work via ML wizard.
   This should need time series related anomaly detection as well.


-- 
============================
Blog: http://srinathsview.blogspot.com twitter:@srinath_perera
Site: http://people.apache.org/~hemapani/
Photos: http://www.flickr.com/photos/hemapani/
Phone: 0772360902
_______________________________________________
Architecture mailing list
[email protected]
https://mail.wso2.org/cgi-bin/mailman/listinfo/architecture

Reply via email to