A recommendation engine for your applications codemotion ams
3 likes453 views
The document discusses the development of a recommendation engine, highlighting the differences between content-based and collaborative filtering methods, along with their limitations such as cold starts and sparsity. It provides guidance on constructing user and item models, utilizing similarity metrics to recommend items based on user preferences, and outlines practical implementation steps, including event handling and data modeling. Additionally, the document includes useful resources and links for further exploration of recommendation systems.
![Item cold start
Cannot predict ratings for new item till some
similar users have rated it [No problem for
content-based]](https://image.slidesharecdn.com/arecommendationengineforyourapplicationscodemotionams-170518081454/85/A-recommendation-engine-for-your-applications-codemotion-ams-27-320.jpg)
![$ pio app new MyApp1
[INFO] [App$] Initialized Event Store for this app ID: 1.
[INFO] [App$] Created new app:
[INFO] [App$] Name: MyApp1
[INFO] [App$] ID: 1
[INFO] [App$] Access Key:
3mZWDzci2D5YsqAnqNnXH9SB6Rg3dsTBs8iHkK6X2i54IQsIZI1eEeQQyMfs7b3F
$ pio eventserver](https://image.slidesharecdn.com/arecommendationengineforyourapplicationscodemotionams-170518081454/85/A-recommendation-engine-for-your-applications-codemotion-ams-45-320.jpg)
![engine.json
"datasource": {
"params" : {
"appName": “MyApp1”,
"eventNames": [“buy”, “view”]
}
},](https://image.slidesharecdn.com/arecommendationengineforyourapplicationscodemotionams-170518081454/85/A-recommendation-engine-for-your-applications-codemotion-ams-57-320.jpg)
A recommendation engine for your applications codemotion ams
- 1. A recommendation engine for your applications Michele Orselli | ideato AMSTERDAM 16 - 17 MAY 2017
- 15. Hotel A Hotel B Hotel C John 3 5 Jane 3 Fred 1 0 Tom 4 AVG 3.5 3 0
- 16. Content based
- 17. User rate items We build a model of user preference Look for similar items based on the model
- 18. Action 0.7 Sci Fi 3.2 Vin Diesel 1.2 … … Title Star Trek Genre Sci Fi Actors William Shatner LeonardNimoy User Model Title XXX Genre Action Actors Vin Diesel Asia Argento Title Kick Ass Genre Action Actors Nicholas Cage Mark Strong
- 20. Need to know items content User cold start: time to learn important features for the user What if user interest change? Lack of serendipity: accidentally discover something you like
- 22. User-User Select people of my neighbourhood with similar taste. If other people share my taste I want their opinion combined
- 23. Item-Item Find an items where I have expressed an opinion and look how other people felt about it
- 25. Sparsity When recommending from a large item set, users will have rated only some of the items
- 26. User Cold start Not enough known about new user to decide who is similar
- 27. Item cold start Cannot predict ratings for new item till some similar users have rated it [No problem for content-based]
- 28. Scalability With millions of ratings, computations become slow
- 29. An example
- 30. Item1 Item2 Item3 Item4 Item5 Joe 8 1 ? 2 7 Tom 2 ? 5 7 5 Alice 5 4 7 4 7 Bob 7 1 7 3 8 How similar are Joe and Tom? How similar are Joe and Bob?
- 31. Only consider items both users have rated For each item - Compute difference in the users’ ratings - Take the average of this difference over the items Item1 Item2 Item3 Item4 Item5 Joe 8 1 ? 2 7 Tom 2 ? 5 7 5 Alice 5 4 7 4 7 Bob 7 1 7 3 8
- 32. Sim(Joe, Tom) = (|8-2| + |2-7| + |7-5|)/3 = 13/3 = 4.3 Sim(Joe, Alice) = (|8-5| + |1-4| + |2-4| + |7-7|)/4 = 2 Sim(Joe, Bob) = (|8-7| + |1-1| + |2-3| + |7-8|)/4 = 0.75 Item1 Item2 Item3 Item4 Item5 Joe 8 1 ? 2 7 Tom 2 ? 5 7 5 Alice 5 4 7 4 7 Bob 7 1 7 3 8
- 33. Similarity Bob 1.57 Alice 0.33 Tom 0.18 D = 1 / 1 + d
- 34. Recommend what similar user have rated highly To calculate rating of an item to recommend, give weight to each user’s recommendations based on how similar they are to you.
- 35. Rating(Joe, Item3) = (1.57 * 7 + 0.33 * 7 + 0.18 * 5) / 3 10.99 + 2.31 + 0.9 / 3 = 4.3 Similarity Bob 1.57 Alice 0.33 Tom 0.18
- 36. use entire matrix or use a K-nn algorithm: people who historically have the same tastes as me aggregate using weighted sum weights depends on similarity
- 37. Our domain
- 38. Domain: online book shop, both paper and digital Recommend titles, old and new - Who bought this also bought - You might like
- 39. PredictionIO
- 42. The event server
- 44. Pattern: user -- action -- item User 1 purchased product X User 2 viewed product Y User 1 added product Z in the cart
- 45. $ pio app new MyApp1 [INFO] [App$] Initialized Event Store for this app ID: 1. [INFO] [App$] Created new app: [INFO] [App$] Name: MyApp1 [INFO] [App$] ID: 1 [INFO] [App$] Access Key: 3mZWDzci2D5YsqAnqNnXH9SB6Rg3dsTBs8iHkK6X2i54IQsIZI1eEeQQyMfs7b3F $ pio eventserver
- 46. Server runs on port 7070 by default $ curl -i -X GET http://localhost:7070 {“status":"alive"}
- 47. $ curl -i -X GET “http://localhost: 7070/events.json? accessKey=$ACCESS_KEY"
- 48. Events modeling what can/should we model? rate, like, buy, view, depending on the algorithm
- 52. setUser($uid, array $properties=array(), $eventTime=null) unsetUser($uid, array $properties, $eventTime=null) deleteUser($uid, $eventTime=null) setItem($iid, array $properties=array(), $eventTime=null) unsetItem($iid, array $properties, $eventTime=null) deleteItem($iid, $eventTime=null)
- 53. recordUserActionOnItem($event, $uid, $iid, array $properties=array(), $eventTime=null) createEvent(array $data) getEvent($eventId)
- 54. Engines
- 56. $ pio template get apache/incubator- predictionio-template-recommender MyRecommendation $ cd MyRecommendation
- 57. engine.json "datasource": { "params" : { "appName": “MyApp1”, "eventNames": [“buy”, “view”] } },
- 58. $ pio build —verbose $ pio train $ pio deploy
- 63. Implementation
- 65. 2 kind of suggestions - who bought this also bought (recommendation) - you may like (similarities)
- 66. View Like (add to basket, add to wishlist) Conversion (buy) Recorded in batch
- 67. 4 engines 2 for books, 2 for ebooks (not needed now) Retrained every night with new data
- 68. recordLike($user, array $item) recordConversion($user, array $item) recordView($user, array $item) createUser($uid)
- 69. getRecommendation($uid, $itype, $n=self::N_SUGGESTION) getSimilarity($iid, $itype, $n=self::N_SUGGESTION)
- 71. Cold start if we don’t get enough suggestion switch to non personalised (also for non logged users)
- 74. The end?
- 75. There’s more...
- 81. Do it on your own https://github.com/grahamjenson/ger
- 84. Michele Orselli CTO@Ideato _orso_ micheleorselli / ideatosrl mo@ideato.it