AUTOMATED MEASUREMENT OF CONTENT QUALITY

• US 20190310989A1
• Filed: 06/21/2019
• Published: 10/10/2019
• Est. Priority Date: 04/06/2017
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Golland, David, Huang, Eric, Chase, Patrick, Patry, Alexandre, Sinha, Shakti Dhirendraji
• Granted Patent: US 11,188,545 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 16/9024 Graphs; Linked lists G06F16/9027 takes precedence
• CPC Class Codes: H04L 51/52 for supporting social networking services
• CPC Class Codes: H04L 67/306 User profiles
• CPC Class Codes: H04L 67/535 Tracking the activity of the user network monitoring arrangements H04L43/00; recording of computer activity G06F11/34

Abstract

A system and method for calculating quality score for digital content are provided. In example embodiments, a first graph is generated comprising a user node and an article node, the user node corresponds to a user and the article node corresponds to an article. An edge is generated between the user node and the article node in the first graph based on a first action. A second graph is generated comprising the user node and the article node. An edge is generated between the user node and the article node in the second graph based on a second action type. A first authority score is calculated for the article node within the first graph. A second authority score is calculated for the article node within the second graph. A quality score is calculated for the article.

Claims

1. A system comprising:

a processor, and a memory including instructions, which when executed by the processor, cause the processor to;

generate a first graph comprising a user node and a text snippet node, the user node uniquely corresponds to a user and the text snippet node uniquely corresponds to a text snippet;

generate an edge between the user node and the text snippet node in the first graph based on a first action type by the user in association with the text snippet;

generate a second graph comprising the user node and the text snippet node;

generate an edge between the user node and the text snippet node in the second graph based on a second action type by the user in association with the text snippet;

calculate a first authority score for the text snippet node within the first graph based on a hub score of the user node in the first graph;

calculate a second authority score for the text snippet node within the second graph based on a hub score of the user node in the second graph; and

calculate a quality score for the text snippet based on a weighted average of the first authority score and the second authority score; and

display the text snippet in a graphical user interface based on the quality score.

2. The system of claim 1, further comprising:

determine a weight for each node based on a logistic regression model using data collected from user engagement with text snippets within a browsing session; and

the weighted average is based on the determined weight for each node.

3. The system of claim 1, wherein:

the first graph comprises a plurality of user nodes;

generate edges between the plurality of user nodes and the text snippet node within the first graph; and

the calculating the first authority score based on a normalized sum of the hubs scores corresponding to the plurality of the user nodes.

4. The system of claim 1, wherein:

the graph includes a second user node, the second user node corresponds to a second user; and

generate an edge between the user node and the second user node based on an action by the user in association with the second user.

5. The system of claim 1, wherein:

the graph includes a second text snippet node, the second text snippet node corresponds to a second text snippet; and

generate an edge between the text snippet node and the second text snippet node based on a hyperlink within the text snippet that corresponds to the second text snippet.

6. The system of claim 1, wherein:

the calculating the first authority score for the text snippet node within the first graph and the second authority score within the second graph is based on a stochastic approach for linked-structure analysis.

7. The system of claim 1, wherein:

the first action type is an action selected from a group consisting of sharing, commenting, liking, dwelling, or clicking on the text snippet.

8. The system of claim 1, wherein:

the first graph comprises a plurality of text snippet nodes;

determine that a text snippet node within the plurality of text snippet nodes has no connecting edges; and

calculate a quality score for the text snippet node within the plurality of text snippet nodes based on an authority score corresponding to the web domain associated with the text snippet node.

9. A method comprising:

using one or more computer processors;

generating a first graph comprising a user node and a text snippet node, the user node uniquely corresponds to a user and the text snippet node uniquely corresponds to a text snippet;

generating an edge between the user node and the text snippet node in the first graph based on a first action type by the user in association with the text snippet;

generating a second graph comprising the user node and the text snippet node;

generating an edge between the user node and the text snippet node in the second graph based on a second action type by the user in association with the text snippet;

calculating a first authority score for the text snippet node within the first graph based on a hub score of the user node in the first graph;

calculating a second authority score for the text snippet node within the second graph based on a hub score of the user node in the second graph; and

calculating a quality score for the text snippet based on a weighted average of the first authority score and the second authority score; and

displaying the text snippet in a graphical user interface based on the quality score.

10. The method of claim 9, further comprising:

determining a weight for each node based on a logistic regression model using data collected from user engagement with text snippets within a browsing session; and

the weighted average is based on the determined weight for each node.

11. The method of claim 9, wherein:

the first graph comprises a plurality of user nodes;

generating edges between the plurality of user nodes and the text snippet node within the first graph; and

the calculating the first authority score based on a normalized sum of the hubs scores corresponding to the plurality of the user nodes.

12. The method of claim 9, wherein:

the first graph includes a second user node, the second user node corresponds to a second user; and

generating an edge between the user node and the second user node based on an action by the user in association with the second user.

13. The method of claim 9, wherein:

the first graph includes a second text snippet node, the second text snippet node corresponds to a second text snippet; and

generating an edge between the text snippet node and the second text snippet node based on a hyperlink within the text snippet that corresponds to the second text snippet.

14. The method of claim 9, wherein:

the calculating the first authority score for the text snippet node within the first graph and the second authority score within the second graph is based on a stochastic approach for linked-structure analysis.

15. The method of claim 9, wherein:

the first action type is an action selected from a group consisting of sharing, commenting, liking, dwelling, or clicking on the text snippet.

16. The method of claim 9, wherein

the first graph comprises a plurality of text snippet nodes;

determine that a text snippet node within the plurality of text snippet nodes has no connecting edges; and

calculating a quality score for the text snippet node within the plurality of text snippet nodes based on an authority score corresponding to the web domain associated with the text snippet node.

17. A machine-readable medium not having any transitory signals and storing instructions that, when executed by at least one processor of a machine, cause the machine to perform operations comprising:

generating a first graph comprising a user node and a text snippet node, the user node uniquely corresponds to a user and the text snippet node uniquely corresponds to a text snippet;

generating an edge between the user node and the text snippet node in the first graph based on a first action type by the user in association with the text snippet;

generating a second graph comprising the user node and the text snippet node;

generating an edge between the user node and the text snippet node in the second graph based on a second action type by the user in association with the text snippet;

calculating a first authority score for the text snippet node within the first graph based on a hub score of the user node in the first graph;

calculating a second authority score for the text snippet node within the second graph based on a hub score of the user node in the second graph; and

calculating a quality score for the text snippet based on a weighted average of the first authority score and the second authority score; and

displaying the text snippet in a graphical user interface based on the quality score.

18. The machine-readable medium of claim 17, wherein the operations further comprise:

determining a weight for each node based on a logistic regression model using data collected from user engagement with text snippets within a browsing session; and

the weighted average is based on the determined weight for each node.

19. The machine-readable medium of claim 17, wherein:

the first graph includes a second user node, the second user node corresponds to a second user; and

wherein the operations further comprise generating an edge between the user node and the second user node based on an action by the user in association with the second user.

20. The machine-readable medium of claim 17, wherein:

the first graph comprises a plurality of text snippet nodes;

determine that a text snippet node within the plurality of text snippet nodes has no connecting edges; and

wherein the operations further comprise calculating a quality score for the text snippet node within the plurality of text snippet nodes based on an authority score corresponding to the web domain associated with the text snippet node.