Correlated information recommendation

• US 8,380,784 B2
• Filed: 07/20/2009
• Issued: 02/19/2013
• Est. Priority Date: 07/24/2008
• Status: Active Grant

• Current Assignee: Alibaba Group Holding Ltd.
• Original Assignee: Alibaba Group Holding Ltd.
• Inventors: Zhu, Baojin, Zhang, Qing
• Primary Examiner(s): Chan, Wing F
• Assistant Examiner(s): WOO, ANDREW M
• Application Number:
  US12/663,243
• Publication Number:
  US 20110238730A1
• Time in Patent Office: 1,310 Days
• Field of Search: 709/203, 709/201
• US Class Current: 709/203
• CPC Class Codes: G06F 16/2425 Iterative querying; Query formulation based on the results of a preceding query
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 16/2465 Query processing support for facilitating data mining operations in structured databases
• 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 16/337 Profile generation, learning or modification
• CPC Class Codes: G06F 16/90335 Query processing
• CPC Class Codes: G06F 16/907 Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
• CPC Class Codes: G06F 16/908 using metadata automatically derived from the content
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 16/9536 Search customisation based on social or collaborative filtering
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: H04L 67/00 Network arrangements or protocols for supporting network services or applications user-to-user messaging H04L51/00; network arrangements, protocols or services for supporting real-time applications in data packet communications networks H04L65/00
• CPC Class Codes: H04L 67/01 Protocols

Abstract

Method and apparatus for information recommendation are provided. In one aspect, a method for information recommendation uses correlated information combinations to improve recommendation accuracy. Upon receiving data from a client indicating a visited information type, the method obtains correlated information types related to the visited information type from stored records. The correlated information types providing one or more correlated information combinations each including at least two correlated information types. For each of the one or more correlated information combinations, the method computes a degree of correlation between the correlated information combination and the visited information type. The method selects a target correlated information combination with a satisfying degree of correlation, and recommends the target correlated information combination to the client.

Claims

1. A method for information recommendation, the method comprising:

receiving data from a client indicating a visited information type;

obtaining a plurality of correlated information types related to the visited information type from stored records, the plurality of correlated information types providing one or more correlated information combinations each including at least two correlated information types;

for each of the one or more correlated information combinations, computing a degree of correlation between the correlated information combination and the visited information type, the computing comprising;

selecting the one or more correlated information combinations based on an application scenario of the client;

determining an occurrence property of each of the one or more correlated information combinations; and

computing the degree of correlation between each of the one or more correlated information combinations and the visited information type using the occurrence property; and

selecting from the one or more correlated information combinations a target correlated information combination whose degree of correlation satisfies a requirement, and recommending the target correlated information combination to the client.

2. The method as recited in claim 1, wherein the occurrence property comprises one or more of a frequency of occurrence, a time of occurrence, or a platform of occurrence.

3. The method as recited in claim 1, wherein the visited information type includes one or more of merchandise information, blog information, group information, post information, product information, news information, message information, keyword information, or advertisement information.

4. The method as recited in claim 1, wherein obtaining the plurality of correlated information types related to the visited information type comprises:

obtaining from the stored records one or more other clients which have used the visited information type; and

obtaining one or more other information types which have been used by the one or more other clients from the stored records.

5. The method as recited in claim 1, further comprising:

obtaining user information of the client and user information of one or more other clients which have used the visited information type; and

wherein computing the degree of correlation between each of the one or more correlated information combinations and the visited information type further comprises;

determining a weight coefficient for each of the one or more correlated information combinations according to the user information of the client and the user information of the one or more other clients which have used the visited information type; and

computing the degree of correlation between the correlated information combination and the visited information type based on the occurrence property of each of the one or more correlated information combinations and the weight coefficient.

6. The method as recited in claim 5, wherein the user information of the client and the user information of the one or more other clients each include one or more of user age information, user education level, user career information, and user family information.

7. The method as recited in claim 1, wherein selecting from the one or more correlated information combinations a target correlated information combination comprises:

ranking the one or more correlated information combinations according to the respective degree of correlation, and selecting a correlated information combination whose degree of correlation is ranked among the highest to be the target correlated information combination.

8. The method as recited in claim 1, wherein selecting from the one or more correlated information combinations a target correlated information combination comprises:

comparing the degree of correlation of each of the one or more correlated information combinations with a set value, and selecting a correlated information combination whose degree of correlation is greater than the set value to be the target correlated information combination.

9. The method as recited in claim 1, further comprising:

converting the target correlated information combination into a hypertext format; and

sending the hypertext format to the client.

10. An apparatus for information recommendation, the apparatus comprising:

one or more processors;

memory;

a receiving unit stored in the memory and executable by the one or more processors that is configured to;

obtain from stored records a plurality of correlated information types related to a visited information type indicated by a client, the plurality of correlated information types providing one or more correlated information combinations each including at least two correlated information types, and

obtain user information of the client and user information of one or more other clients which have used the visited information type and obtaining one or more other information types which have been used by the one or more other clients from the stored records;

a computation unit stored in the memory and executable by the one or more processors that is configured to;

obtain an occurrence property of each of the one or more correlated information combinations,

select the one or more correlated information combinations of at least two correlated information types based on an application scenario of the client,

determine a weight coefficient for each of the one or more correlated information combinations according to the user information of the client and the user information of the one or more other clients which have used the visited information type, and

compute a degree of correlation between each of the one or more correlated information combinations and the visited information type based on an occurrence property of the respective correlated information combination and the weight coefficient; and

a recommendation unit stored in the memory and executable by the one or more processors that is configured to select from the one or more correlated information combinations a target correlated information combination whose degree of correlation satisfies a requirement, and recommend the target correlated information combination to the client.

11. The apparatus as recited in claim 10, wherein the apparatus further comprises:

a storage unit configured to store the stored records containing information that have been used by the client and one or more other clients.

12. The apparatus as recited in claim 10, wherein the apparatus comprises a server computer.

13. A method comprising:

under control of one or more processors configured with executable instructions;

receiving, from a client, data indicating a visited information type, the visited information type comprising a type of information that is visited or used by a user of the client;

obtaining records of one or more other clients that have visited or used the visited information type;

obtaining a plurality of correlated information types that have been visited or used by the one or more other clients;

selecting a plurality of combinations of correlated information from the plurality of correlated information based on an application scenario of the client, the application scenario indicating which information type of the plurality of correlated information is undesirable to the user and which information type of the plurality of correlated information is desirable to the user;

computing a degree of correlation between each combination of correlated information and the visited information type;

selecting, from the plurality of combinations of correlated information, one or more target combinations of correlated information having a degree of correlation that satisfies a requirement; and

recommending the one or more target combinations of correlated information to the user through the client.

14. The method as recited in claim 13, wherein the plurality of correlated information types comprises at least two of:

merchandise information, blog information, group information, post information, product information, news information, message information, keyword information and advertisement information.

15. The method as recited in claim 13, further comprising obtaining an occurrence property of each combination of correlated information, and wherein computing the degree of correlation comprises computing the degree of correlation between each combination of correlated information and the visited information type based on the occurrence property of respective combination of correlated information.

16. The method as recited in claim 15, wherein the occurrence property of each combination of correlated information comprises a frequency of occurrence, a time of occurrence and a place of occurrence of the respective combination of correlated information.