Building a Topical Learning Model in a Content Management System

• US 20160147891A1
• Filed: 11/25/2014
• Published: 05/26/2016
• Est. Priority Date: 11/25/2014
• Status: Active Grant

• Current Assignee: Chegg Incorporated
• Original Assignee: Chegg Incorporated
• Inventors: Chhichhia, Charmy, Le Chevalier, Vincent
• Granted Patent: US 9,852,132 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/244 Grouping and aggregation
• CPC Class Codes: G06F 16/285 Clustering or classification
• 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 16/93 Document management systems
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 16/986 Document structures and storage, e.g. HTML extensions

Abstract

A content management system receives a plurality of topics extracted from documents stored by the system. Pairings between the topics are generated, where a pairing between two topics is generated responsive to the two topics appearing in proximity to one another in one or more of the documents. A complexity of each received topic is also determined. The content management system generates a progression of the topics based on the complexity of the topics and the pairings between the topics. The progression comprises a sequential ordering of paired topics, in which a topic in the ordering has a higher complexity than a preceding topic. Responsive to a user of the content management system accessing content associated with a topic in the progression, a next topic in the progression is recommended to the user.

Claims

1. A method for recommending a topic to a user of a content management system, the method comprising:

receiving a plurality of topics extracted from documents of the content management system;

generating pairings between the received topics, a pairing between two topics generated responsive to the two topics appearing in proximity to one another in one or more of the documents;

determining a complexity of each received topic;

generating a progression of the topics based on the complexity of the topics and the pairings between the topics, the progression comprising a sequential ordering of paired topics in which a topic in the ordering has a higher complexity than a preceding topic; and

responsive to a user of the content management system accessing content associated with a topic in the progression, recommending a next topic in the progression to the user.

2. The method of claim 1, wherein generating the pairings between the received topics comprises:

identifying two topics appearing in proximity to one another in one or more of the documents;

scoring the two topics based on a degree of correlation between the two topics; and

pairing the two topics responsive to the score being greater than a threshold.

3. The method of claim 1, wherein determining the complexity of each received topic comprises, for a received topic:

generating a complexity measure for each document from which the topic was extracted;

identifying a normalized location of the topic in each document; and

determining the complexity of the topic based on the complexity measure for each document and the normalized location of the topic in each document.

4. The method of claim 1, wherein the progression of topics is generated from a plurality of sequences of topics, and wherein selecting the sequences of topics to include in the progression comprises:

generating a plurality of candidate sequences of the paired topics, each candidate sequence including two or more topics ordered according to the complexities of the two or more topics;

determining a frequency of each candidate sequence in the documents of the content management system; and

responsive to the frequency of a candidate sequence exceeding a threshold, selecting the candidate sequence for inclusion in the progression of topics.

5. The method of claim 1, wherein the documents of the content management system include documents having a plurality of different publication dates, the method further comprising:

generating a growth rate for each topic based on a frequency of the topic in documents published at different times;

wherein the progression of topics is further generated based on the growth rate of the topics.

6. The method of claim 5, wherein generating the progression of topics based on the growth rate of the topics comprises:

for a first topic in the progression, identifying a plurality of topics paired to the first topic that have a higher complexity than the first topic; and

selecting from the identified topics, a second topic to succeed the first topic in the progression based on the growth rate of the second topic, the second topic having a higher growth rate than other identified topics.

7. The method of claim 1, wherein recommending the next topic in the progression to the user comprises:

displaying a portion of a document including the next topic to the user.

8. The method of claim 1, further comprising:

displaying a representation of the progression of topics to the user, each topic in the representation selectable by the user to access a document from which the selected topic was extracted.

9. A non-transitory computer readable storage medium storing computer program instructions for recommending a topic to a user of a content management system, the computer program instructions when executed by a processor causing the processor to:

receive a plurality of topics extracted from documents of the content management system;

generate pairings between the received topics, a pairing between two topics generated responsive to the two topics appearing in proximity to one another in one or more of the documents;

determine a complexity of each received topic;

generate a progression of the topics based on the complexity of the topics and the pairings between the topics, the progression comprising a sequential ordering of paired topics in which a topic in the ordering has a higher complexity than a preceding topic; and

responsive to a user of the content management system accessing content associated with a topic in the progression, recommend a next topic in the progression to the user.

10. The non-transitory computer readable storage medium of claim 9, wherein the computer program instructions causing the processor to generate the pairings between the received topics comprise instructions that when executed by the processor cause the processor to:

identify two topics appearing in proximity to one another in one or more of the documents;

score the two topics based on a degree of correlation between the two topics; and

pair the two topics responsive to the score being greater than a threshold.

11. The non-transitory computer readable storage medium of claim 9, wherein the computer program instructions causing the processor to determine the complexity of each received topic comprise instructions that when executed by the processor cause the processor to, for a received topic:

generate a complexity measure for each document from which the topic was extracted;

identify a normalized location of the topic in each document; and

determine the complexity of the topic based on the complexity measure for each document and the normalized location of the topic in each document.

12. The non-transitory computer readable storage medium of claim 9, wherein the progression of topics is generated from a plurality of sequences of topics, and wherein the computer program instructions causing the processor to select the sequences of topics to include in the progression instructions that when executed by the processor cause the processor to:

generate a plurality of candidate sequences of the paired topics, each candidate sequence including two or more topics ordered according to the complexities of the two or more topics;

determine a frequency of each candidate sequence in the documents of the content management system; and

responsive to the frequency of a candidate sequence exceeding a threshold, select the candidate sequence for inclusion in the progression of topics.

13. The non-transitory computer readable storage medium of claim 9, wherein the documents of the content management system include documents having a plurality of different publication dates, further comprising computer program instructions causing the processor to:

generate a growth rate for each topic based on a frequency of the topic in documents published at different times;

wherein the progression of topics is further generated based on the growth rate of the topics.

14. The non-transitory computer readable storage medium of claim 13, wherein the computer program instructions causing the processor to generate the progression of topics based on the growth rate of the topics comprise computer program instructions that when executed by the processor cause the processor to:

for a first topic in the progression, identify a plurality of topics paired to the first topic that have a higher complexity than the first topic; and

select from the identified topics, a second topic to succeed the first topic in the progression based on the growth rate of the second topic, the second topic having a higher growth rate than other identified topics.

15. The non-transitory computer readable storage medium of claim 9, wherein the computer program instructions causing the processor to recommend the next topic in the progression to the user comprise computer program instructions that when executed by the processor cause the processor to:

display a portion of a document including the next topic to the user.

16. The non-transitory computer readable storage medium of claim 9, further comprising computer program instructions that when executed by the processor cause the processor to:

display a representation of the progression of topics to the user, each topic in the representation selectable by the user to access a document from which the selected topic was extracted.