Recommendation method and device

• US 10,298,705 B2
• Filed: 11/15/2016
• Issued: 05/21/2019
• Est. Priority Date: 11/17/2015
• Status: Active Grant

• Current Assignee: Alibaba Group Holding Ltd.
• Original Assignee: Alibaba Group Holding Ltd.
• Inventors: Liao, Chuang, Yang, Chao, Ni, Na, Liu, Zhongyi
• Primary Examiner(s): Miah, Razu A
• Application Number:
  US15/352,310
• Publication Number:
  US 20170142209A1
• Time in Patent Office: 917 Days
• CPC Class Codes: H04L 65/40 Support for services or applications
• CPC Class Codes: H04L 67/306 User profiles
• CPC Class Codes: H04L 67/51 Discovery or management thereof, e.g. service location protocol [SLP] or web services
• CPC Class Codes: H04L 67/535 Tracking the activity of the user network monitoring arrangements H04L43/00; recording of computer activity G06F11/34

Abstract

Disclosed is a recommendation method and device. The method comprises acquiring a user'"'"'s Internet service sequence, a user preference vector for Internet services, and an attribute vector for Internet services identified in the user'"'"'s Internet service sequence each of the foregoing based on user network behavior data generated when the user accesses the Internet services; calculating user preference scores for one or more candidate Internet service providers, wherein the user preferences scores are calculated based on the user'"'"'s Internet service sequence, the user preference vector for the Internet services, the attribute vector for the Internet services, and Internet service lists provided by the candidate Internet service providers; and sending a recommendation of potential users to the candidate Internet service providers based on the user preference scores for the candidate Internet service providers. The present disclosure can recommend more high-quality potential users to Internet service providers, to increase user subscription.

Claims

1. A method for recommending users to an Internet service provider, the method comprising:

retrieving, by a processor, a log of user behavior data with Internet services provided by Internet service providers using a network system, the user behavior data comprising network interactions received from users of the Internet services;

building, by the processor, a user'"'"'s Internet service sequence based on the log of user behavior by identifying a sequence of Internet services included in the log during a predefined time period, the Internet service sequence organized as a time sequence of network behavior occurrences;

building, by the processor, a user preference vector for Internet services based on executing vectorized coding on user preference data associated with the Internet services, the user preference data comprising an amount of interactions with each Internet service, and

building, by the processor, an attribute vector for Internet services identified in the user'"'"'s Internet service sequence by executing vectorized coding on attribute information associated with each of the Internet services identified in the user'"'"'s Internet service sequence;

calculating, by the processor, user preference scores for one or more candidate Internet service providers, wherein the user preferences scores are calculated based on the user'"'"'s Internet service sequence, the user preference vector for the Internet services, the attribute vector for the Internet services, and Internet service lists provided by the candidate Internet service providers; and

sending a recommendation, by the processor, of potential users to the candidate Internet service providers based on the user preference scores for the candidate Internet service providers.

2. The method according to claim 1, wherein executing vectorized coding on the attribute information associated with each of the Internet service identified in the user'"'"'s Internet service sequence comprises:

maximizing value of formula

3. The method according to claim 1, wherein executing vectorized coding on the user preference data associated with the Internet services comprises:

maximizing value of formula

4. The method according to claim 1, wherein calculating user preference scores for one or more Internet service providers, wherein the user preferences scores are calculated based on the user'"'"'s Internet service sequence, the user preference vector for the Internet services, the attribute vector for the Internet services, and the Internet service lists provided by the candidate Internet service providers comprises:

calculating the user preference score for a candidate Internet service provider based on the formula

5. The method according to claim 1, wherein sending a recommendation of potential users to candidate Internet service providers based on the user preference scores for the Internet service providers comprises:

normalizing the user preference scores for the Internet service providers based on user preference scores for all Internet service providers, in order to acquire standardized user preference scores for the candidate Internet service providers; and

sending a recommendation of potential users to the candidate Internet service providers based on the standardized user preference scores for the candidate Internet service providers.

6. The method according to claim 5, wherein sending a recommendation of potential users to candidate Internet service providers according to the standardized user preference scores comprises:

determining user activeness within a predetermined time range; and

sending a recommendation of the potential users to the candidate Internet service providers based on the user activeness within the predetermined time range and the standardized user preference scores for the candidate Internet service providers.

7. The method according to claim 6, wherein acquiring user activeness within a predetermined time range comprises determining the user activeness within the predetermined time range based on one or more Internet service detail pages visited by the user within the predetermined time range.

8. The method according to claim 1, wherein the user is one of a plurality of users comprising users who have not previously been associated with or have not had any direct network behavior associated with the Internet service providers.

9. A recommendation device comprising:

a processor;

a storage medium for tangibly storing thereon program logic for execution by the processor, the program logic comprising;

vector acquisition logic executed by the processor for

retrieving a log of user behavior data with Internet services provided by Internet service providers using a network system, the user behavior data comprising network interactions received from users of the Internet services,

building a user'"'"'s Internet service sequence based on the log of user behavior by identifying a sequence of Internet services included in the log during a predefined time period, the Internet service sequence organized as a time sequence of network behavior occurrences,

building a user preference vector for Internet services based on executing vectorized coding on user preference data associated with the Internet services, the user preference data comprising an amount of interactions with each Internet service, and

building an attribute vector for Internet services identified in the user'"'"'s Internet service sequence by executing vectorized coding on attribute information associated with each of the Internet services identified in the user'"'"'s Internet service sequence;

score calculation logic executed by the processor for calculating user preference scores for one or more candidate Internet service providers, wherein the user preferences scores are calculated based on the user'"'"'s Internet service sequence, the user preference vector for the Internet services, and the attribute vector for the Internet services, and Internet service lists provided by the candidate Internet service providers; and

recommendation logic executed by the processor for sending a recommendation of potential users to the candidate Internet service providers based on the user preference scores for the candidate Internet service providers.

10. The device according to claim 9, wherein the vector acquisition logic further comprises:

logic for maximizing the value of formula

11. The device according to claim 9, wherein the vector further comprises:

logic for maximizing the value of formula

12. The device according to claim 9, wherein the score acquisition logic further comprises:

user preference score calculating logic executed by the processor for calculating the user preference score for a candidate Internet service provider based on the formula

13. The device according to claim 9, wherein the recommendation logic further comprises:

normalizing logic executed by the processor for normalizing the user preference scores for the Internet service providers based on the user preference scores for all Internet service providers, in order to acquire standardized user preference scores for the candidate Internet service providers;

potential user recommending logic executed by the processor for sending a recommendation of potential users to the candidate Internet service providers based on the standardized user preference scores for the candidate Internet service providers.

14. The device according to claim 13, wherein the recommendation logic further comprises:

user activeness acquiring logic executed by the processor for determining user activeness within a predetermined time range;

wherein the potential user recommending logic further comprises logic for sending a recommendation of the potential users to the candidate Internet service providers based on the user activeness within the predetermined time range and the standardized user preference scores for the candidate Internet service providers.

15. The device according to claim 14, wherein the user activeness acquiring logic further comprises detail page determining logic executed by the processor for determining the user activeness within the predetermined time range based on a number of Internet service detail pages visited by the user within the predetermined time range.

16. The device according to claim 9, wherein the user is one of a plurality of users comprising users who have not previously been associated or have had any direct network behavior dealings with the Internet service providers.

17. A non-transitory computer readable storage medium tangibly storing computer program instructions capable of being executed by a computer processor, the computer program instructions defining the steps of:

retrieving, by a processor, a log of user behavior data with Internet services provided by Internet service providers using a network system, the user behavior data comprising network interactions received from users of the Internet services;

building, by the processor, a user'"'"'s Internet service sequence based on the log of user behavior by identifying a sequence of Internet services included in the log during a predefined time period, the Internet service sequence organized as a time sequence of network behavior occurrences;

building, by the processor, a user preference vector for Internet services based on executing vectorized coding on user preference data associated with the Internet services, the user preference data comprising an amount of interactions with each Internet service;

building, by the processor, an attribute vector for Internet services identified in the user'"'"'s Internet service sequence by executing vectorized coding on attribute information associated with each of the Internet services identified in the user'"'"'s Internet service sequence;

calculating, by the processor, user preference scores for one or more candidate Internet service providers, wherein the user preferences scores are calculated based on the user'"'"'s Internet service sequence, the user preference vector for the Internet services, the attribute vector for the Internet services, and Internet service lists provided by the candidate Internet service providers; and

sending a recommendation, by the processor, of potential users to candidate Internet service providers based on the user preference scores for the candidate Internet service providers.