Recommending media programs based on media program popularity

1. A system comprising:

one or more computer processors; and

one or more non-transitory computer readable devices that include instructions that, when executed by the one or more computer processors, causes the processors to perform operations, the operations comprising:

determining, by a computer system, a plurality of data sources that relate to a plurality of media programs, for which descriptions of the plurality of media programs are unavailable;

training, a recommendation engine, by applying a machine learning technique to identify concepts from the plurality of data sources, wherein the machine learning technique comprises analyzing the plurality of data sources that relate to the plurality of media programs for which the descriptions of the plurality of media programs are unavailable, to derive one or more keywords of each of the plurality of media programs;

receiving, by the computer system, information expressing a user's interest in one or more media programs of the plurality of media programs;

determining, by the computer system, commonalities between the user's interest in one or more media programs and the one or more keywords of each of the plurality of media programs;

identifying, by the computer system using the recommendation engine, one or more other media programs among the plurality of media programs based on the commonalities;

identifying, by the computer system, a completion of a content display, on a user interface, of at least one media program of the plurality of media programs for which the descriptions are unavailable; and

generating, by the recommendation engine for presenting on the user interface, a recommendation of at least one of the one or more other media programs.

2. A computer-implemented method, comprising:

determining, by a computer system, a plurality of data sources that relate to a plurality of media programs, for which descriptions of the plurality of media programs are unavailable;

training, a recommendation engine, by applying a machine learning technique to identify concepts from the plurality of data sources, wherein the machine learning technique comprises analyzing the plurality of data sources that relate to the plurality of media programs for which the descriptions of the plurality of media programs are unavailable, to derive one or more keywords of each of the plurality of media programs;

receiving, by the computer system, information expressing a user's interest in one or more media programs of the plurality of media programs;

determining, by the computer system, commonalities between the user's interest in one or more media programs and the one or more keywords of each of the plurality of media programs;

identifying, by the computer system using the recommendation engine, one or more other media programs among the plurality of media programs based on the commonalities;

identifying, by the computer system, a completion of a content display, on a user interface, of at least one media program of the plurality of media programs for which the descriptions are unavailable; and

generating, by the recommendation engine for presenting on the user interface, a recommendation of at least one of the one or more other media programs.

3. The method of claim 2 , wherein the data sources comprise closed caption data for one or more media programs among the plurality of media programs.

4. The method of claim 2 , wherein the data sources comprise at least one of blogs relating to one or more media programs among the plurality of media programs, or web site content that describes one or more media programs among the plurality of media programs.

5. The method of claim 2 , further comprising deriving the information expressing a user's interest in one or more media programs based on the user's web activity data.

6. The method of claim 5 , wherein the user's web activity data includes web-click data derived from media-related search queries submitted by the user.

7. The method of claim 2 , wherein the plurality of data sources that relate to the plurality of media programs offset a lack of program description available from one or more electronic program guide data providers.

8. The method of claim 2 , wherein training a recommendation engine further comprises analyzing audience measurement data relating to the plurality of media programs, wherein identifying the one or more media programs is further based on a popularity of the one or more media programs, wherein the popularity of the one or more media programs is determined from analysis of the audience measurement data relating to the plurality of media programs.

9. The method of claim 2 , wherein training a recommendation engine further comprises identifying at least one program among the plurality of media programs as new.

10. The method of claim 9 , wherein identifying the at least one program among the plurality of media programs as new comprises determining that less than a threshold amount of click data relating to the at least one program is available.

11. The method of claim 9 , wherein identifying the at least one program among the plurality of media programs as new comprises determining that less than a threshold amount of audience measurement data relating to the at least one program is available.

12. A non-transitory computer-readable medium storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to perform operations comprising:

determining, by a computer system, a plurality of data sources that relate to a plurality of media programs, for which descriptions of the plurality of media programs are unavailable;

training, a recommendation engine, by applying a machine learning technique to identify concepts from the plurality of data sources, wherein the machine learning technique comprises analyzing the plurality of data sources that relate to the plurality of media programs for which the descriptions of the plurality of media programs are unavailable, to derive one or more keywords of each of the plurality of media programs;

receiving, by the computer system, information expressing a user's interest in one or more media programs of the plurality of media programs;

determining, by the computer system, commonalities between the user's interest in one or more media programs and the one or more keywords of each of the plurality of media programs;

identifying, by the computer system using the recommendation engine, one or more other media programs among the plurality of media programs based on the commonalities;

identifying, by the computer system, a completion of a content display, on a user interface, of at least one media program of the plurality of media programs for which the descriptions are unavailable; and

generating, by the recommendation engine for presenting on the user interface, a recommendation of at least one of the one or more other media programs.

13. The non-transitory computer-readable medium of claim 12 , wherein the data sources comprise one of closed caption data for one or more media programs among the plurality of media programs.

14. The non-transitory computer-readable medium of claim 12 , wherein the data sources comprise at least one of blogs relating to one or more media programs among the plurality of media programs, or web site content that describes one or more media programs among the plurality of media programs.

15. The non-transitory computer-readable medium of claim 12 , further comprising deriving the information expressing a user's interest in one or more media programs based on the user's web activity data.

16. The non-transitory computer-readable medium of claim 15 , wherein the user's web activity data includes web-click data derived from media-related search queries submitted by the user.

17. The non-transitory computer-readable medium of claim 12 , wherein the plurality of data sources that relate to the plurality of media programs offset a lack of program description available from one or more electronic program guide data providers.

18. The non-transitory computer-readable medium of claim 12 , wherein training a recommendation engine further comprises analyzing audience measurement data relating to the plurality of media programs, wherein identifying the one or more media programs is further based on a popularity of the one or more media programs, wherein the popularity of the one or more media programs is determined from analysis of the audience measurement data relating to the plurality of media programs.

19. The non-transitory computer-readable medium of claim 12 , wherein training a recommendation engine further comprises identifying at least one program among the plurality of media programs as new.

20. The non-transitory computer-readable medium of claim 19 , wherein identifying the at least one program among the plurality of media programs as new comprises determining that less than a threshold amount of click data relating to the at least one program is available.