classification template

11 views
Skip to first unread message

YOUSSEF ASSATA

unread,
Apr 12, 2017, 10:16:39 AM4/12/17
to action...@googlegroups.com
hello everyone please can you suggest me a classification template or a tuto based on items and their probabilities
thanks cordially 

Pat Ferrel

unread,
Apr 12, 2017, 11:35:17 AM4/12/17
to YOUSSEF ASSATA, actionml-user, us...@predictionio.incubator.apache.org
We support the Universal Recommender here. See the PredictionIO mailing list for further support (CCed)


On Apr 12, 2017, at 7:16 AM, YOUSSEF ASSATA <assata...@gmail.com> wrote:

hello everyone please can you suggest me a classification template or a tuto based on items and their probabilities
thanks cordially 

--
You received this message because you are subscribed to the Google Groups "actionml-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to actionml-use...@googlegroups.com.
To post to this group, send email to action...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/actionml-user/CAFm3dtC6krZOQMceEopOifozF0Vqc1RV7AQXF9k0YWSbP1hk_A%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

Pat Ferrel

unread,
Apr 12, 2017, 12:11:38 PM4/12/17
to YOUSSEF ASSATA, actionml-user
The UR can recommend anything you have usage data for. Meaning anything you can tie to a user.

To recommend topics, you would need to somehow record a user’s preference for topics (user-id, topic-pref, topic-id) then record other user preference indicators as secondary events, name these events in the engine.json in the correct order with topic-pref first like this “eventNames”: [“topic-pref”, “read”, …]  Now when you ask for a user’s recs you send a user-id, or for similar topics, send “item”: “topic-id” in the query but you always get back topics if “topic-pref” is the first event name.

Notice that to recommend a “read” you would create a model with the “read” indicator first, if we want to recommend anything, it should be first.

The UR actually represents a landscape of algorithms that can use all user usage data to recommend different things based on how you configure it.



On Apr 12, 2017, at 8:55 AM, YOUSSEF ASSATA <assata...@gmail.com> wrote:

i think yo didn't get my question  : i'm looking for a template that make recommandation based on a database with topics and their probability => so i can make a topic in a specific categorie 
thank

2017-04-12 16:35 GMT+01:00 Pat Ferrel <p...@occamsmachete.com>:
We support the Universal Recommender here. See the PredictionIO mailing list for further support (CCed)

On Apr 12, 2017, at 7:16 AM, YOUSSEF ASSATA <assata...@gmail.com> wrote:

hello everyone please can you suggest me a classification template or a tuto based on items and their probabilities
thanks cordially 

--
You received this message because you are subscribed to the Google Groups "actionml-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to actionml-user+unsubscribe@googlegroups.com.
Reply all
Reply to author
Forward
0 new messages