Approach for estimating user ratings of items

• US 8,271,338 B2
• Filed: 09/18/2009
• Issued: 09/18/2012
• Est. Priority Date: 02/06/2004
• Status: Active Grant

• Current Assignee: Netflix, Inc.
• Original Assignee: Netflix, Inc.
• Inventors: Hunt, Neil Duncan, Lanning, Stanley Miguel, Hastings, W. Reed, Purcell, Shawn Michael, Shih, Lawrence Wen-Kai, Ciancutti, John Robert
• Primary Examiner(s): ZURITA, JAMES H
• Application Number:
  US12/562,741
• Publication Number:
  US 20100010877A1
• Time in Patent Office: 1,096 Days
• Field of Search: 705/26, 705/27
• US Class Current: 705/26
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0201 Market modelling; Market analysis; Collecting market data
• CPC Class Codes: G06Q 30/0204 Market segmentation
• CPC Class Codes: G06Q 30/0217 involving input on products or services in exchange for incentives or rewards
• CPC Class Codes: G06Q 30/06 Buying, selling or leasing transactions
• CPC Class Codes: G06Q 30/0623 Item investigation
• CPC Class Codes: G06Q 40/04 Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange

Abstract

According to an embodiment of the invention, an approach is provided for estimating how a particular user would rate a particular item from a plurality of items. The approach is applicable to any type of items, including rental items such as movies, music and games, and the invention is not limited to any particular type of item. One or more items from the plurality of items that have ratings similar to the particular item are identified. One or more other users are identified that have given ratings to the one or more items that are substantially similar to ratings given by the particular user to the one or more items. An estimation is made of how the particular user would rate the particular item based upon ratings for the particular item given by the one or more other users.

Claims

1. A method for generating an estimation of how a particular user would rate a particular item that the particular user has not yet rated, the method comprising:

identifying one or more users that have given ratings to one or more similar items that are substantially similar to ratings given by the particular user to the one or more similar items; and

generating an estimation of how the particular user would rate the particular item that the particular user has not yet rated based upon ratings given to the particular item by the one or more users;

wherein the method is performed by one or more computing devices.

2. The method of claim 1, further comprising:

identifying the one or more similar items by identifying one or more items that have been given ratings similar to ratings given to the particular item that the particular user has not yet rated.

3. The method of claim 1, further comprising:

identifying the one or more similar items by identifying one or more items that are similar to the particular item that the particular user has not yet rated based on semantic information available for the one or more items and the particular item.

4. The method of claim 1, further comprising:

selecting one or more reference users based upon selection criteria; and

wherein the one or more users are identified from the one or more reference users.

5. The method of claim 4, wherein the selection criteria comprise a minimum number of ratings and wherein selecting one or more reference users comprises selecting one or more reference users that have each given at least the minimum number of ratings.

6. The method of claim 4, wherein the selection criteria comprise a maximum number of ratings and wherein selecting one or more reference users comprises selecting one or more reference users that have each given at most the maximum number of ratings.

7. The method of claim 4, wherein the selection criteria comprise an average rating range and wherein selecting one or more reference users comprises selecting one or more reference users that have each given ratings, an average of which is within the average rating range.

8. The method of claim 4, wherein the selection criteria comprise a tolerance of a normal distribution and wherein selecting one or more reference users comprises selecting one or more reference users that have each given ratings, a distribution of which is within the tolerance.

9. The method of claim 4, wherein the selection criteria comprise a range of a standard deviation and wherein selecting one or more reference users comprises selecting one or more reference users that have each given ratings, a standard deviation of which is within the range.

10. The method of claim 1, further comprising:

maintaining ratings data that represents ratings given to a plurality of items by a plurality of users;

wherein the plurality of items includes the one or more similar items and the particular item; and

wherein the plurality of users includes the particular user.

11. The method of claim 10, wherein the ratings data is maintained in an array data structure in a volatile memory of a computing device, the array having a plurality of cells, each cell corresponding to a single item of the plurality of items and a single user of the plurality of users.

12. The method of claim 11, wherein at least one cell of the array comprises data representing a rating given by the user corresponding to the cell to the item corresponding to the cell and at least one other cell of the array comprises data representing an estimation of how the user corresponding to the other cell would rate the item corresponding to the other cell.

13. A computer-readable storage medium storing computer-executable instructions for generating an estimation of how a particular user would rate a particular item that the particular user has not yet rated, the instructions, when executed by one or more computing devices, cause the one or more computing devices to perform:

identifying one or more users that have given ratings to one or more similar items that are substantially similar to ratings given by the particular user to the one or more similar items; and

generating an estimation of how the particular user would rate the particular item that the particular user has not yet rated based upon ratings given to the particular item by the one or more users.

14. The computer-readable storage medium of claim 13, further storing instructions which, when executed by the one or more computing devices, cause the one or more computing devices to perform:

identifying the one or more similar items by identifying one or more items that have been given ratings similar to ratings given to the particular item that the particular user has not yet rated.

15. The computer-readable storage medium of claim 13, further storing instructions which, when executed by the one or more computing devices, cause the one or more computing devices to perform:

identifying the one or more similar items by identifying one or more items that are similar to the particular item that the particular user has not yet rated based on semantic information available for the one or more items and the particular item.

16. The computer-readable storage medium of claim 13, further storing instructions which, when executed by the one or more computing devices, cause the one or more computing devices to perform:

selecting one or more reference users based upon selection criteria; and

wherein the one or more users are identified from the one or more reference users.

17. The computer-readable storage medium of claim 16, wherein the selection criteria comprise a minimum number of ratings and wherein selecting one or more reference users comprises selecting one or more reference users that have each given at least the minimum number of ratings.

18. The computer-readable storage medium of claim 16, wherein the selection criteria comprise a maximum number of ratings and wherein selecting one or more reference users comprises selecting one or more reference users that have each given at most the maximum number of ratings.

19. The computer-readable storage medium of claim 16, wherein the selection criteria comprise an average rating range and wherein selecting one or more reference users comprises selecting one or more reference users that have each given ratings, an average of which is within the average rating range.

20. The computer-readable storage medium of claim 16, wherein the selection criteria comprise a tolerance of a normal distribution and wherein selecting one or more reference users comprises selecting one or more reference users that have each given ratings, a distribution of which is within the tolerance.

21. The computer-readable storage medium of claim 16, wherein the selection criteria comprise a range of a standard deviation and wherein selecting one or more reference users comprises selecting one or more reference users that have each given ratings, a standard deviation of which is within the range.

22. The computer-readable storage medium of claim 13, further storing instructions which, when executed by the one or more computing devices, cause the one or more computing devices to perform:

maintaining ratings data that represents ratings given to a plurality of items by a plurality of users;

wherein the plurality of items includes the one or more similar items and the particular item; and

wherein the plurality of users includes the particular user.

23. The computer-readable storage medium of claim 22, wherein the ratings data is maintained in an array data structure in a volatile memory of a computing device, the array having a plurality of cells, each cell corresponding to a single item of the plurality of items and a single user of the plurality of users.

24. The computer-readable storage medium of claim 23, wherein at least one cell of the array comprises data representing a rating given by the user corresponding to the cell to the item corresponding to the cell and at least one other cell of the array comprises data representing an estimation of how the user corresponding to the other cell would rate the item corresponding to the other cell.