The UR is supported here: https://groups.google.com/forum/#!forum/actionml-user <https://groups.google.com/forum/#!forum/actionml-user>
First 0.2.3 has a bug in how it weights biased items so you’ll want to get 0.3.0, or 0.4.0 coming next week. This will require you to upgrade the ActionML version of PIO 0.9.7-aml here: actionml.com/docs/install When there is an Apache release we’ll support it instead but that’s a week or 2 away. However even with your versions: 1) Filters and boosts work with item properties not “correlators” like the “search" values. 2) all correlators are theoretically weighted equally, there is no method provided to boost certain correlators over others and this would be dangerous in theory. 3) just because a user has certain secondary events, does not mean these events pass the correlation test an so the items may not have all secondary events associated, in fact that would be bad because some are random non-correlating actions. 4) put another way, your query does nothing with “search” and would not be able to parse it in any case. See the docs here for query specs: http://actionml.com/docs/ur_queries <http://actionml.com/docs/ur_queries> The user history of all events is used by simply specifying the user-id so there is no reason to do this in the query. Maybe we should backup and discuss what you are trying to do with this? On Sep 18, 2016, at 11:03 AM, Malay Tripathi <malaytripat...@gmail.com> wrote: Hi, Will very much appreciate any response! I am using "purchase" as primary event & "search term" as secondary event in below set of data. (Only data i uploaded in engine & trained)- I understand- Item 2851561,2929427 and 2351393 are higher score because both user 101 & 102 purchased same item 2562459. (primary event relation) Item 2748084,2859418 follows because they were purchased by user 103 and both user 102 & 103 searched for"BELTS BAG” (secondary event relation) Followed by 289818 & 2900542 with score 0 because of no relation. BUT I was expecting that if I use search term “NINE WEST NOLYN” with user 102 then it would increase the score of 289818 & 2900542 move them up in response because user 104 also searched for “NINE WEST NOLYN” and purchased 289818 & 2900542, hence a stronger relationship than user 101 purchased or event relationship with user 103. Recommendation requests- A3837044:bin m828420$ curl -H "Content-Type: application/json" -d '{"user": "102"}' http://localhost:8000/queries.json <http://localhost:8000/queries.json> {"itemScores":[{"item":"2851561","score":0.4863930344581604},{"item":"2929427","score":0.3595937192440033},{"item":"2351393","score":0.16528037190437317},{"item":"2748084","score":0.039376020431518555},{"item":"2859418","score":0.039376020431518555},{"item":"2900542","score":0.0},{"item":"2898183","score":0.0}]} A3837044:bin m828420$ curl -H "Content-Type: application/json" -d '{"user": "102","search":"BELTS BAG"}' http://localhost:8000/queries.json <http://localhost:8000/queries.json> {"itemScores":[{"item":"2851561","score":0.4863930344581604},{"item":"2929427","score":0.3595937192440033},{"item":"2351393","score":0.16528037190437317},{"item":"2748084","score":0.039376020431518555},{"item":"2859418","score":0.039376020431518555},{"item":"2900542","score":0.0},{"item":"2898183","score":0.0}]} A3837044:bin m828420$ curl -H "Content-Type: application/json" -d '{"user": "102","search":"NINE WEST NOLYN"}' http://localhost:8000/queries.jso <http://localhost:8000/queries.jso> {"itemScores":[{"item":"2851561","score":0.4863930344581604},{"item":"2929427","score":0.3595937192440033},{"item":"2351393","score":0.16528037190437317},{"item":"2748084","score":0.039376020431518555},{"item":"2859418","score":0.039376020431518555},{"item":"2900542","score":0.0},{"item":"2898183","score":0.0}]} Data I trained on- {"event":"purchase","entityType":"user","entityId":"101","targetEntityType":"item","targetEntityId":"2562459","properties":{},"eventTime":"2016-05-22T12:00:00.000-06:00"} {"event":"search","entityType":"user","entityId":"101","targetEntityType":"item","targetEntityId":"SHOES","properties":{},"eventTime":"2016-09-04T12:00:00.000-06:00"} {"event":"search","entityType":"user","entityId":"101","targetEntityType":"item","targetEntityId":"BAGS","properties":{},"eventTime":"2016-09-04T12:00:00.000-06:00"} {"event":"purchase","entityType":"user","entityId":"101","targetEntityType":"item","targetEntityId":"2929427","properties":{},"eventTime":"2016-05-22T12:00:00.000-06:00"} {"event":"purchase","entityType":"user","entityId":"101","targetEntityType":"item","targetEntityId":"2351393","properties":{},"eventTime":"2016-05-22T12:00:00.000-06:00"} {"event":"purchase","entityType":"user","entityId":"101","targetEntityType":"item","targetEntityId":"2851561","properties":{},"eventTime":"2016-05-22T12:00:00.000-06:00"} {"event":"purchase","entityType":"user","entityId":"102","targetEntityType":"item","targetEntityId":"2837836","properties":{},"eventTime":"2016-05-22T12:00:00.000-06:00"} {"event":"search","entityType":"user","entityId":"102","targetEntityType":"item","targetEntityId":"SHOES","properties":{},"eventTime":"2016-09-04T12:00:00.000-06:00"} {"event":"purchase","entityType":"user","entityId":"102","targetEntityType":"item","targetEntityId":"2562459","properties":{},"eventTime":"2016-05-22T12:00:00.000-06:00"} {"event":"search","entityType":"user","entityId":"102","targetEntityType":"item","targetEntityId":"BELTS BAG","properties":{},"eventTime":"2016-09-04T12:00:00.000-06:00"} {"event":"purchase","entityType":"user","entityId":"103","targetEntityType":"item","targetEntityId":"2748084","properties":{},"eventTime":"2016-05-22T12:00:00.000-06:00"} {"event":"search","entityType":"user","entityId":"103","targetEntityType":"item","targetEntityId":"BELTS BAG","properties":{},"eventTime":"2016-09-04T12:00:00.000-06:00"} {"event":"purchase","entityType":"user","entityId":"103","targetEntityType":"item","targetEntityId":"2859418","properties":{},"eventTime":"2016-05-22T12:00:00.000-06:00"} {"event":"search","entityType":"user","entityId":"103","targetEntityType":"item","targetEntityId":"BAG","properties":{},"eventTime":"2016-09-04T12:00:00.000-06:00"} {"event":"purchase","entityType":"user","entityId":"104","targetEntityType":"item","targetEntityId":"2898183","properties":{},"eventTime":"2016-05-22T12:00:00.000-06:00"} {"event":"search","entityType":"user","entityId":"104","targetEntityType":"item","targetEntityId":"BAG","properties":{},"eventTime":"2016-09-04T12:00:00.000-06:00"} {"event":"purchase","entityType":"user","entityId":"104","targetEntityType":"item","targetEntityId":"2900542","properties":{},"eventTime":"2016-05-22T12:00:00.000-06:00"} {"event":"search","entityType":"user","entityId":"104","targetEntityType":"item","targetEntityId":"NINE WEST NOLYN","properties":{},"eventTime":"2016-09-06T12:00:00.000-06:00"} <engine.json>
