Collaborative filtering

• US 9,177,048 B1
• Filed: 03/26/2013
• Issued: 11/03/2015
• Est. Priority Date: 02/16/2007
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Das, Abhinandan S., Garg, Ashutosh, Datar, Mayur Dhondu
• Primary Examiner(s): Spieler, William
• Application Number:
  US13/851,092
• Time in Patent Office: 952 Days
• Field of Search: 707/732
• US Class Current: 1/1
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 16/335 Filtering based on additional data, e.g. user or group profiles filtering in web context G06F16/9535, G06F16/9536
• CPC Class Codes: G06F 16/90335 Query processing

Abstract

Systems, methods, and apparatus, including computer program products, for collaborative filtering are provided. A method is provided. The method includes clustering a plurality of entities with respect to one or more latent variables in a probability distribution model of a relationship between a set of entities and a set of items, the probability distribution model comprising a probability distribution of the set of items with respect to the latent variables. The method also includes, as new items are added to the set of items, updating the probability distribution of the set of the items with respect to the latent variables, and generating an updated relationship score for an entity with respect to the set of items based on the entity'"'"'s fractional membership in the clustering with respect to the latent variables and based on the updated probability distribution of the set of the items with respect to the latent variables.

Claims

1. A computer-implemented method comprising:

generating an initial overall probability distribution p(s|u) as a combination of an initial first probability distribution p(z|u) and an initial second probability distribution p(s|z), wherein the initial first probability distribution p(z|u) is a probability of a particular category of a plurality of categories given a user of a set of users, wherein the categories are represented by one or more latent variables, and wherein the initial second probability distribution p(s|z) is a probability distribution of a set of items with respect to the one or more latent variables;

calculating an updated second probability distribution p(s|z)_new of a current set of items with respect to the one or more latent variables including, wherein the updated second probability distribution is calculated using counter values determined based on prior user selections of items in the set of items and according to user membership in the categories represented by the latent variables, wherein each counter value corresponds to a category of which the user is a member, and wherein each counter value is fractionally incremented relative to other categories of which the user is also a member; and

generating a relationship score for each of one or more items in the current set of items, wherein each relationship score generated for a particular item relates the particular item to relating to a particular user in the set of users based on the particular user'"'"'s category memberships and the updated second probability distribution.

2. The method of claim 1, further comprising:

calculating an updated overall probability distribution based on the initial first probability distribution and the updated second probability distribution.

3. The method of claim 1, wherein calculating the updated second probability distribution comprises calculating, for each category of the categories, a fraction of counts for an item in the category relative to all of the counter values for items in the particular category.

4. The method of claim 1, wherein the relationship score is used to generate an item recommendation for a particular user.

5. The method of claim 1, wherein determining the counter values comprises determining the counter values based on one or more of the prior user selections that have each been discounted as a function of time.

6. The method of claim 1, further comprising modifying the relationship score for the particular user based on a history of items selected by the particular user within a period of time.

7. The method of claim 1, wherein calculating the updated second probability distribution of the set of items with respect to the latent variables is determined by:

8. The method of claim 1, wherein a prior selection of an item by a user results in one or more of the counter values being incremented by a total count of 1 distributed as a function of member categories of the user that selected the item.

9. The method of claim 8, wherein the counter values of member categories are incremented according to a first probability distribution relating users and categories.

10. A system comprising:

one or more computer-readable storage media having instructions stored thereon; and

data processing apparatus programmed to execute the instructions to perform operations comprising;

generating an initial overall probability distribution p(s|u) as a combination of an initial first probability distribution p(z|u) and an initial second probability distribution p(s|z), wherein the initial first probability distribution p(z|u) is a probability of a particular category of a plurality of categories given a user of a set of users, wherein the categories are represented by one or more latent variables, and wherein the initial second probability distribution p(s|z) is a probability distribution of a set of items with respect to the one or more latent variables;

calculating an updated second probability distribution p(s|z)_new of a current set of items with respect to the one or more latent variables including, wherein the updated second probability distribution is calculated using counter values determined based on prior user selections of items in the set of items and according to user membership in the categories represented by the latent variables, wherein each counter value corresponds to a category of which the user is a member, and wherein each counter value is fractionally incremented relative to other categories of which the user is also a member; and

generating a relationship score for each of one or more items in the current set of items, wherein each relationship score generated for a particular item relates the particular item to relating to a particular user in the set of users based on the particular user'"'"'s category memberships and the updated second probability distribution.

11. The system of claim 10, wherein the data processing apparatus is further programmed to execute instructions to perform operations comprising:

calculating an updated overall probability distribution based on the initial first probability distribution and the updated second probability distribution.

12. The system of claim 10, wherein calculating the updated second probability distribution comprises calculating, for each category of the categories, a fraction of counts for an item in the category relative to all of the counter values for items in the particular category.

13. The system of claim 10, wherein the relationship score is used to generate an item recommendation for a particular user.

14. The system of claim 10, wherein determining the counter values comprises determining the counter values based on one or more of the prior user selections that have each been discounted as a function of time.

15. The system of claim 10, wherein the data processing apparatus is further programmed to execute instructions to perform operations comprising modifying the relationship score for the particular user based on a history of items selected by the particular user within a period of time.

16. The system of claim 10, wherein calculating the updated second probability distribution of the set of items with respect to the latent variables is determined by:

17. The system of claim 10, wherein a prior selection of an item by a user results in one or more of the counter values being incremented by a total count of 1 distributed as a function of member categories of the user that selected the item.

18. The system of claim 17, wherein the counter values of member categories are incremented according to a first probability distribution relating users and categories.

19. A non-transitory computer-readable medium having instructions stored thereon which, when executed by data processing apparatus, cause the data processing apparatus to perform operations comprising:

generating an initial overall probability distribution p(s|u) as a combination of an initial first probability distribution p(z|u) and an initial second probability distribution p(s|z), wherein the initial first probability distribution p(z|u) is a probability of a particular category of a plurality of categories given a user of a set of users, wherein the categories are represented by one or more latent variables, and wherein the initial second probability distribution p(s|z) is a probability distribution of a set of items with respect to the one or more latent variables;

calculating an updated second probability distribution p(s|z)_new of a current set of items with respect to the one or more latent variables including, wherein the updated second probability distribution is calculated using counter values determined based on prior user selections of items in the set of items and according to user membership in the categories represented by the latent variables, wherein each counter value corresponds to a category of which the user is a member, and wherein each counter value is fractionally incremented relative to other categories of which the user is also a member; and

generating a relationship score for each of one or more items in the current set of items, wherein each relationship score generated for a particular item relates the particular item to relating to a particular user in the set of users based on the particular user'"'"'s category memberships and the updated second probability distribution.