Method and system for content recommendation

1. A computer-executable method, comprising:

selecting a content piece from a content library;

extracting, by a computer using a natural language processing (NLP) technique, one or more keywords from the content piece;

determining a domain associated with the content piece based on the extracted one or more keywords;

obtaining hierarchical domain knowledge of the determined domain, wherein the obtained hierarchical domain knowledge is based on a tree structure comprising a root node corresponding to a domain name, a number of leaf nodes corresponding to feature words each having a respective defined meaning within the determined domain, and one or more branch nodes corresponding to one or more categories within the determined domain;

obtaining one or more domain-knowledge graphs, a domain-knowledge graph of the one or more domain-knowledge graphs including an entity name and a number of attribute words associated with the entity name, the entity name and an attribute word indicating a feature combination word that is specific to the domain;

generating a feature tag for the content piece based on the extracted one or more keywords and the obtained hierarchical domain knowledge;

generating an attribute tag for a user based on historical data associated with the user; and

recommending one or more content pieces from the content library to the user based on feature tags associated with the one or more content pieces and the attribute tag for the user,

wherein the generating the feature tag includes:

comparing the extracted one or more keywords with one or more feature combination words defined by the one or more domain-knowledge graphs;

identifying an extracted keyword that matches a feature combination word of the one or more feature combination words; and

adding the identified extracted keyword to the feature tag.

2. The method of claim 1 , wherein the hierarchical domain knowledge specifies at least: the domain name, a number of categories within the domain, and a number of feature words within each category.

3. The method of claim 2 , wherein generating the feature tag comprises:

determining a category associated with the content piece;

identifying a number of feature words associated with the determined category; and

comparing the extracted one or more keywords with the identified feature words associated with the determined category.

4. The method of claim 3 , wherein generating the feature tag comprises:

identifying an extracted keyword that matches one of the identified feature words associated with the determined category; and

adding the identified extracted keyword to the feature tag.

5. The method of claim 3 , further comprising:

ranking the extracted one or more keywords based on a result of the comparison between the extracted one or more keywords and the identified feature words associated with the determined category; and

adding a number of extracted keywords having rankings within a predetermined ranking range in the feature tag.

6. The method of claim 5 , wherein the extracted one or more keywords are associated with corresponding weight values, and wherein ranking the extracted one or more keywords comprises updating the weight values based on the result of the comparison and ranking the extracted one or more keywords based on the updated weight values.

7. The method of claim 1 , wherein generating the attribute tag for the user comprises:

identifying a second content piece that has been read by the user;

obtaining a second feature tag for the second content piece; and

adding at least a portion of the second feature tag to the attribute tag for the user.

8. The method of claim 1 , wherein obtaining the hierarchical domain knowledge comprises applying a domain-knowledge-based model that has been previously trained using a machine-learning technique.

9. A computer system, comprising:

a processor; and

a storage device coupled to the processor and storing instructions which when executed by the processor cause the processor to perform acts, the acts comprising:

selecting a content piece from a content library;

extracting, by a computer using a natural language processing (NLP) technique, one or more keywords from the content piece;

determining a domain associated with the content piece based on the extracted one or more keywords;

obtaining hierarchical domain knowledge of the determined domain, wherein the obtained hierarchical domain knowledge is based on a tree structure comprising a root node corresponding to a domain name, a number of leaf nodes corresponding to feature words each having a defined meaning within the determined domain, and one or more branch nodes corresponding to one or more categories within the determined domain;

obtaining one or more domain-knowledge graphs, a domain-knowledge graph of the one or more domain-knowledge graphs including an entity name and a number of attribute words associated with the entity name, the entity name and an attribute word indicating a feature combination word that is specific to the domain;

generating a feature tag for the content piece based on the extracted one or more keywords and the obtained hierarchical domain knowledge;

generating an attribute tag for a user based on historical data associated with the user; and

recommending one or more content pieces from the content library to the user based on feature tags associated with the one or more content pieces and the attribute tag for the user,

wherein the generating the feature tag includes:

comparing the extracted one or more keywords with one or more feature combination words defined by the one or more domain-knowledge graphs;

identifying an extracted keyword that matches a feature combination word of the one or more feature combination words; and

adding the identified extracted keyword to the feature tag.

10. The computer system of claim 9 , wherein the hierarchical domain knowledge specifies at least: the domain name, a number of categories within the domain, and a number of feature words within each category.

11. The computer system of claim 10 , wherein generating the feature tag comprises:

determining a category associated with the content piece;

identifying a number of feature words associated with the determined category; and

comparing the extracted one or more keywords with the identified feature words associated with the determined category.

12. The computer system of claim 11 , wherein generating the feature tag comprises:

identifying an extracted keyword that matches one of the identified feature words associated with the determined category; and

adding the identified extracted keyword to the feature tag.

13. The computer system of claim 11 , wherein the acts further comprise:

ranking the extracted one or more keywords based on a result of the comparison between the one or more extracted keywords and the identified feature words associated with the determined category; and

adding a number of extracted keywords having rankings within a predetermined ranking range in the feature tag.

14. The computer system of claim 13 , wherein the extracted one or more keywords are associated with corresponding weight values, and wherein ranking the extracted one or more keywords comprises updating the weight values based on the result of the comparison and ranking the extracted one or more keywords based on the updated weight values.

15. The computer system of claim 9 , wherein generating the attribute tag for the user comprises:

identifying a second content piece that has been read by the user;

obtaining a second feature tag for the second content piece; and

adding at least a portion of the second feature tag to the attribute tag for the user.

16. The computer system of claim 9 , wherein obtaining the hierarchical domain knowledge comprises applying a domain-knowledge-based model that has been previously trained using a machine-learning technique.

17. A non-transitory storage medium having computer executable instructions stored thereon, the computer executable instructions, when executed by a processor, enabling the processor to implement acts comprising:

selecting a content piece from a content library;

extracting, by a computer using a natural language processing (NLP) technique, one or more keywords from the content piece;

determining a domain associated with the content piece based on the extracted one or more keywords;

obtaining hierarchical domain knowledge of the determined domain, wherein the obtained hierarchical domain knowledge is based on a tree structure comprising a root node corresponding to a domain name, a number of leaf nodes corresponding to feature words each having a respective defined meaning within the determined domain, and one or more branch nodes corresponding to one or more categories within the determined domain;

obtaining one or more domain-knowledge graphs, a domain-knowledge graph of the one or more domain-knowledge graphs including an entity name and a number of attribute words associated with the entity name, the entity name and an attribute word indicating a feature combination word that is specific to the domain;

generating a feature tag for the content piece based on the extracted one or more keywords and the obtained hierarchical domain knowledge;

generating an attribute tag for a user based on historical data associated with the user; and

recommending one or more content pieces from the content library to the user based on feature tags associated with the one or more content pieces and the attribute tag for the user,

wherein the generating the feature tag includes:

comparing the extracted one or more keywords with one or more feature combination words defined by the one or more domain-knowledge graphs;

identifying an extracted keyword that matches a feature combination word of the one or more feature combination words; and

adding the identified extracted keyword to the feature tag.

18. The non-transitory storage medium of claim 17 , wherein the hierarchical domain knowledge specifies at least: the domain name, a number of categories within the domain, and a number of feature words within each category.

19. The non-transitory storage medium of claim 18 , wherein generating the feature tag comprises:

determining a category associated with the content piece;

identifying a number of feature words associated with the determined category; and

comparing the extracted one or more keywords with the identified feature words associated with the determined category.

20. The non-transitory storage medium of claim 19 , wherein generating the feature tag comprises:

identifying an extracted keyword that matches one of the identified feature words associated with the determined category; and

adding the identified extracted keyword to the feature tag.