Ranking authors in social media systems
The author ranking technique described herein is a technique to rank authors in social media systems along various dimensions, using a variety of statistical methods for utilizing those dimensions. More particularly, the technique ranks authors in social media systems through a combination of statistical techniques that leverage usage metrics, and social and topical graph characteristics. In various exemplary embodiments, the technique can rank author authority by the following: 1) temporal analysis of link sharing in which authority is computed based on a user's propensity to provide early links to web pages that subsequently become popular; 2) topical authority based on the author's links and content updates in specific topic areas; and 3) popularity and influence based on nodal properties of authors.
1. A process implemented by a computing device, the process comprising:
obtaining data indicating that authors in a social media system provided links to network content to other users of the social media system;
based on statistical analysis of the data, ranking individual authors of the social media system according to their propensity to provide individual links to corresponding network content that becomes popular with the other users of the social media system, the ranking comprising:
defining author scores of the individual authors,
defining document scores of the individual links,
determining corresponding chronological orders in which the individual authors of the social media system provided the individual links to the other users of the social media system,
determining respective popularities of the individual links,
updating the document scores using the author scores, the chronological orders in which the individual links were provided by the individual authors, and the respective popularities of the individual links to generate update document scores, and
updating the author scores using the updated document scores, the chronological orders in which the individual links were provided by the individual authors, and the respective popularities of the individual links to generate updated author scores; and
outputting a ranked list of the individual authors based on the updated author scores.
2. The process of claim 1 , wherein the network content comprises web pages and the individual links comprise Uniform Resource Locator (URL) addresses of the individual web pages.
3. The process of claim 2 , wherein:
the author scores are included in an author score vector, the document scores are included in a document score vector, and the chronological orders and the respective popularities are represented in an adjacency matrix,
the updating the document scores comprises multiplying the author score vector by the adjacency matrix, and
the updating the author scores comprises multiplying the document score vector by the adjacency matrix.
4. The process of claim 3 , further comprising:
initializing each entry in the document score vector to a corresponding domain rank.
5. A computing device comprising:
at least one processing unit; and
one or more computer storage media storing instructions which, when executed by the at least one processing unit, cause the at least one processing unit to:
obtain data indicating the authors in a social media system provided links to network content to other users of the social media system;
rank individual authors of the social media system according to their propensity to provide individual links to corresponding network content that becomes popular with the other users of the social media system, the individual authors being ranked by:
defining author scores of the individual authors,
defining document scores of the individual links,
determining corresponding chronological orders in which the individual authors of the social media system provided the individual links to the users of the social media system,
determining the respective popularities of the individual links,
updating the document scores using the author scores, the chronological orders in which the individual links were provided by the individual authors, and the respective popularities of the individual links to generate updated document scores, and
updating the author scores using the updated document scores, the chronological orders in which the individual links were provided by the individual authors, and the respective popularities of the individual links to generate updated author scores; and
output a ranked list of the individual authors based on the updated author scores.
6. The computing device of claim 5 , wherein the instructions cause the at least one processing unit to:
perform multiple iterations of the updating the document scores and the updating the author scores before the ranked list is output.
7. The computing device of claim 5 , wherein the instructions cause the at least one processing unit to:
continue updating the document scores and the author scores until a given condition is reached, wherein the ranked list is output responsive to the given condition being reached.
8. The computing device of claim 7 , wherein the given condition comprises a minimal difference between the updated author scores over consecutive iterations.
9. The computing device of claim 5 , wherein the instructions cause the at least one processing unit to:
update the author scores using a mathematical function having a first component that represents the corresponding chronological orders in an exponent and a second component that applies a logarithm operation to the respective popularities.
10. The computing device of claim 9 , wherein the instructions cause the at least one processing unit to:
update the document scores using the mathematical function having the first component and the second component.
11. The computing device of claim 5 , wherein the document scores represent domain rankings.
12. The computing device of claim 5 , wherein the instructions cause the at least one processing unit to:
rank search results based on the updated author scores.
13. The computing device of claim 5 , wherein the instructions cause the at least one processing unit to:
filter search results to exclude spammers based on the updated author scores.
14. One or more hardware computer storage media storing instructions which, when executed by the at least one processing unit, cause the at least one processing unit to perform acts comprising:
obtaining data indicating that authors in a social media system provided links to network content to other users of the social media system;
ranking individual authors of the social media system according to their propensity to provide individual links to corresponding network content that becomes popular with the other users of the social media system based on statistical analysis of the data, the ranking comprising:
defining author scores of the individual authors,
defining document scores of the individual links
determining corresponding chronological orders in which the individual authors of the social media system provided the individual links to the other users of the social media system,
determining respective popularities of the individual links,
updating the document scores using the author scores, the chronological orders in which the individual links were provided by the individual authors, and the respective popularities of the individual links to generate update document scores, and
updating the author scores using the updated document scores, the chronological orders in which the individual links were provided by the individual authors, and the respective popularities of the individual links to generate updated author scores; and
outputting a ranked list of the individual authors based on the updated author scores.
15. The one or more hardware computer storage media of claim 14 , the acts further comprising:
initializing the author scores to uniform values for each of the individual authors.
16. The one or more hardware computer storage media of claim 14 , wherein the ranked list is sorted by the updated author scores.
17. The one or more hardware computer storage media of claim 14 , the acts further comprising:
updating the document scores and the author scores iteratively using a particular mathematical term.
18. The one or more hardware computer storage media of claim 17 , wherein the particular mathematical term comprises a first component representing the corresponding chronological orders, and a second component representing the respective popularities.
19. The one or more hardware computer storage media of claim 18 , the acts further comprising:
updating the document scores by multiplying the updated author scores by the particular mathematical term; and
updating the author scores by multiplying the updated document scores by the particular mathematical term.
20. The one or more hardware computer storage media of claim 19 , the acts further comprising:
normalizing the updated author scores and the updated document scores, wherein the ranked list is based on the normalized updated author scores and the normalized updated document scores.