Taste profile attributes

• US 10,540,385 B2
• Filed: 03/14/2014
• Issued: 01/21/2020
• Est. Priority Date: 03/15/2013
• Status: Active Grant

• Current Assignee: Spotify AB
• Original Assignee: Spotify AB
• Inventors: Whitman, Brian, Lamere, Paul
• Primary Examiner(s): Khong, Alexander
• Assistant Examiner(s): Morris, John J
• Application Number:
  US14/214,026
• Publication Number:
  US 20140279817A1
• Time in Patent Office: 2,139 Days
• Field of Search: 707742
• CPC Class Codes: G06F 16/437 Administration of user profiles, e.g. generation, initialisation, adaptation, distribution
• CPC Class Codes: G06F 16/637 Administration of user profiles, e.g. generation, initialization, adaptation or distribution

Abstract

Methods, systems and computer program products are provided for summarizing user activity associated with media content by accessing a taste profile containing a representation of media content activity corresponding to at least one of a plurality of items, generating at least one statistic corresponding to the media content activity, and generating a taste profile attribute by using the at least one statistic.

Claims

1. A system for summarizing user activity associated with media content, comprising:

a processor configured to;

access a taste profile associated with a first user containing a plurality of entries defining a representation of media content activity, each entry being associated to at least one corresponding item of a plurality of items, and at least two of the entries of the plurality of entries being action variables defining actions relating to the at least one corresponding item, a first of the actions being playback of the corresponding item associated with the first user, a second of the actions being playback of the corresponding item associated with a plurality of second users other than the first user;

retrieve at least one term from a plurality of term-weight pairs contained in a document, the at least one term being (i) indexed to a corresponding entry of the plurality of entries in the taste profile and (ii) weighted by a weight indicating the closeness of the term to the corresponding entry;

generate statistics corresponding to the media content activity based in part by combining the action variables and one or more of the plurality of term-weight pairs, the statistics including;

i) an affinity statistic; and

ii) a familiarity statistic, the affinity statistic being, for each of the plurality of items, based at least in part on counting a number of playbacks of the item associated with the first user, and the familiarity statistic being, for each of the plurality of items, based at least n part on calculating a probability that the plurality of second users other than the first user know of the item; and

generate taste profile attributes by using at least one of the statistics, the taste profile attributes including;

i) a passion attribute; and

ii) an adventurousness attribute, the passion attribute being based on the affinity statistic and the familiarity statistic and being generated in part by measuring an amount of passion of the first user for each of the plurality of items, a value of the passion attribute being inversely proportional to a value of the familiarity statistic, and the adventurousness attribute being based on the affinity statistic and the familiarity statistic and being generated in part by measuring an amount of adventurousness of the first user;

wherein generating each of the taste profile attributes includes calculating an average of the corresponding one or more statistics across the plurality of items.

2. The system according to claim 1, wherein the probability is further based on any one of, or a combination of, a mention of one or more of the plurality of items on a webpage, and an external source of popularity data.

3. The system according to claim 1, wherein the processor is further operable to generate a diversity attribute, the diversity attribute being based on the affinity statistic and being generated in part by measuring a stylistic distance between the plurality of items, the processor being further operable to, when generating the diversity attribute:

generate a plurality of style maps S_a; and

generate an inverted overlap matrix based on the plurality of style maps,

O<sup>−

1</sup>(S₁,S₂. . . ,S_a),

wherein each of the plurality of style maps S_a is based on a value representative of how connected a predetermined taste profile activity is to a corresponding item of the plurality of items.

4. The system according to claim 1, wherein the adventurousness attribute (a) is generated according to the equation:

5. The system according to claim 1, wherein the processor is further configured to:

calculate a weighted average of an amount representative of a trending popularity across the named collection of items.

6. The system according to claim 1,

wherein the statistics include a recency statistic, the recency statistic being a plurality of scores representative of recency of each of the plurality of items, correspondingly, and wherein the taste profile attributes include an attribute generated by calculating an average of the plurality of scores.

7. The system according to claim 1, wherein at least one of the statistics is based on an age of each of the plurality of items.

8. A non-transitory computer-readable medium having stored thereon one or more sequences of instructions for causing one or more processors to perform:

accessing a taste profile associated with a first user containing a plurality of entries defining a representation of media content activity, each entry being associated to at least one corresponding item of a plurality of items, and at least two of the entries of the plurality of entries being action variables defining actions relating to the at least one corresponding item, a first of the actions being playback of the corresponding item associated with the first user, a second of the actions being playback of the corresponding item associated with a plurality of second users other than the first user;

retrieving at least one term from a plurality of term-weight pairs contained in a document, the at least one term being (i) indexed to a corresponding entry of the plurality of entries in the taste profile and (ii) weighted by a weight indicating the closeness of the term to the corresponding entry;

generating statistics corresponding to the media content activity based in part by combining the action variables and one or more of the plurality of term-weight pairs, the statistics including;

i) an affinity statistic; and

ii) a familiarity statistic, the affinity statistic being, for each of the plurality of items, based at least in part on counting a number of playbacks of the item associated with the first user, and the familiarity statistic being, for each of the plurality of items, based at least in part on calculating a probability that the plurality of second users other than the first user know of the item; and

generating taste profile attributes by using at least one of the statistics, the taste profile attributes including;

i) a passion attribute; and

ii) an adventurousness attribute, the passion attribute being based on the affinity statistic and the familiarity statistic and being generated in part by measuring an amount of passion of the first user for each of the plurality of items, a value of the passion attribute being inversely proportional to a value of the familiarity statistic, and the adventurousness attribute being based on the affinity statistic and the familiarity statistic and being generated in part by measuring an amount of adventurousness of the first user;

wherein generating each of the taste profile attributes includes calculating an average of the corresponding one or more statistics across the plurality of items.

9. The computer-readable medium of claim 8, wherein the probability is further based on any one of, or a combination of, usage data, a mention of one or more of the plurality of items on a webpage, and an external source of popularity data.

10. The computer-readable medium of claim 8, wherein the one or more processors are further operable to generate a diversity attribute, the diversity attribute being based on the affinity statistic and being generated in part by measuring a stylistic distance between the plurality of items, the one or more processors being further operable to, when generating the diversity attribute:

generate a plurality of style maps S_a; and

generate an inverted overlap matrix based on the plurality of style maps,

O<sup>−

1</sup>(S₁,S₂ . . . ,S_a),

wherein each of the plurality of style maps S_a is based on a value representative of how connected a predetermined taste profile activity is to a corresponding item of the plurality of items.

11. The computer-readable medium of claim 8, wherein the adventurousness attribute (a) is generated according to the equation:

12. The computer-readable medium of claim 8, wherein the one or more processors are further configured to:

calculate a weighted average of an amount representative of a trending popularity across the named collection of items.

13. The computer-readable medium of claim 8, wherein the statistics include a recency statistic, the recency statistic being a plurality of scores representative of recency of each of the plurality of items, correspondingly, and wherein the taste profile attributes include an attribute generated by calculating an average of the plurality of scores.

14. The computer-readable medium of claim 8, wherein at least one of the statistics is based on an age of each of the plurality of items.

15. A method for summarizing user activity associated with media content, comprising the steps of:

accessing a taste profile associated with a first user containing a plurality of entries defining a representation of media content activity, each entry being associated to at least one corresponding item of a plurality of items, and at least some of the entries of the plurality of entries being action variables defining actions relating to the at least one corresponding item, a first of the actions being playback of the corresponding item associated with the first user, a second of the actions being playback of the corresponding item associated with a plurality of second users other than the first user;

retrieving at least one term from a plurality of term-weight pairs contained in a document, the at least one term being (i) indexed to a corresponding entry of the plurality of entries in the taste profile and (ii) weighted by a weight indicating the closeness of the term to the corresponding entry;

generating statistics corresponding to the media content activity based in part by combining the action variables and one or more of the plurality of term-weight pairs, the statistics including;

i) an affinity statistic; and

ii) a familiarity statistic, the affinity statistic being, for each of the plurality of items, based at least in part on counting a number of playbacks of the item associated with the first user, and the familiarity statistic being, for each of the plurality of items, based at least in part on calculating a probability that the plurality of second users other than the first user know of the item; and

generating taste profile attributes by using at least one of the statistics, the taste profile attributes including at least;

i) a passion attribute; and

ii) an adventurousness attribute, the passion attribute being based on the affinity statistic and the familiarity statistic and being generated in part by measuring an amount of passion of the first user for each of the plurality of items, a value of the passion attribute being inversely proportional to a value of the familiarity statistic, and the adventurousness attribute being based on the affinity statistic and the familiarity statistic and being generated by measuring an amount of adventurousness of the first user;

wherein generating each of the taste profile attributes includes calculating an average of the corresponding one or more statistics across the plurality of items.

16. The method of claim 15, wherein the probability is further based on any one of, or a of, usage data, a mention of one or more of the plurality of items on a webpage, and an external source of popularity data.

17. The method of claim 15, further comprising generating a diversity attribute, the diversity attribute being based on the affinity statistic and being generated in part by measuring a stylistic distance between the plurality of items, the method further including, when generating the diversity attribute:

generating a plurality of style maps S_a; and

generating an inverted overlap matrix based on the plurality of style maps,

O<sup>−

1</sup>(S₁,S₂ . . . ,S_a),

wherein each of the plurality of style maps S_a is based on a value representative of how connected a predetermined taste profile activity s to a corresponding item of the plurality of items.

18. The method of claim 15, wherein the adventurousness attribute (a) is generated according to the equation:

19. The method of claim 15, further comprising:

calculating a weighted average of an amount representative of a trending popularity cross the named collection of items.

20. The method of claim 15, wherein the statistics include a recency statistic, the recency statistic being a plurality of scores representative of recency of each of the plurality of items, correspondingly, and wherein the one or more taste profile attributes include an attribute generated by calculating an average of the plurality of scores.

21. The method of claim 15, wherein at least one of the statistics is based on an age of each of the plurality of items.