Recommending features for content planning based on advertiser polling and historical audience measurements

1. A method, comprising the steps of:

obtaining historical data comprising content preferences for a plurality of existing content items of an audience having audience segments, wherein said content preferences indicate a performance metric for each content item and audience segment pair, wherein said plurality of existing content items comprise a plurality of content features indicating characteristics of a corresponding content item;

obtaining, for each of a plurality of advertisers, a weight for each of a plurality of said audience segments, wherein a given weight indicates a future interest of a given advertiser in a given audience segment;

calculating, using at least one processing device, a pairwise similarity between a vector of said content preferences and a vector of said weights for said plurality of said audience segments to obtain a ranked list of said plurality of existing content items sorted by said pairwise similarity; and

generating, using said at least one processing device, a summarization of said content features, wherein said summarization of said content features is used to plan one or more new content items based on said ranked list, wherein said summarization of said content features is based on a frequency of the content features in the ranked list of said plurality of existing content items, wherein said one or more new content items are produced after said generating and comprise one or more content features from said summarization.

2. The method of claim 1 , wherein said historical data is organized by one or more of said audience segments.

3. The method of claim 1 , wherein said performance metric comprises an average rating measured for a particular content item.

4. The method of claim 1 , wherein said weight is based on amount of money a given advertiser expects to spend on targeting each of said plurality of audience segments.

5. The method of claim 1 , wherein said pairwise similarity is based on an inverse of a Euclidean distance.

6. The method of claim 1 , further comprising the step of computing a normalization ratio to scale said vector of said weights for said plurality of said audience segments with said performance metrics.

7. The method of claim 1 , wherein said obtaining of said weight for said plurality of said audience segments considers one or more subsets of said plurality of advertisers with similar profiles.

8. A computer program product, comprising a non-transitory machine-readable storage medium having encoded therein executable code of one or more software programs, wherein the one or more software programs when executed by at least one processing device perform the following steps:

obtaining historical data comprising content preferences for a plurality of existing content items of an audience having audience segments, wherein said content preferences indicate a performance metric for each content item and audience segment pair, wherein said plurality of existing content items comprise a plurality of content features indicating characteristics of a corresponding content item;

obtaining, for each of a plurality of advertisers, a weight for each of a plurality of said audience segments, wherein a given weight indicates a future interest of a given advertiser in a given audience segment;

calculating, using said at least one processing device, a pairwise similarity between a vector of said content preferences and a vector of said weights for said plurality of said audience segments to obtain a ranked list of said plurality of existing content items sorted by said pairwise similarity; and

generating, using said at least one processing device, a summarization of said content features, wherein said summarization of said content features is used to plan one or more new content items based on said ranked list, wherein said summarization of said content features is based on a frequency of the content features in the ranked list of said plurality of existing content items, wherein said one or more new content items are produced after said generating and comprise one or more content features from said summarization.

9. The computer program product of claim 8 , wherein said performance metric comprises an average rating measured for a particular content item.

10. The computer program product of claim 8 , wherein said weight is based on amount of money a given advertiser expects to spend on targeting each of said plurality of audience segments.

11. The computer program product of claim 8 , further comprising the step of computing a normalization ratio to scale said vector of said weights for said plurality of said audience segments with said performance metrics.

12. The computer program product of claim 8 , wherein said obtaining of said weight for said plurality of said audience segments considers one or more subsets of said plurality of advertisers with similar profiles.

13. A system, comprising:

a memory; and

at least one processing device, coupled to the memory, operative to implement the following steps:

obtaining historical data comprising content preferences for a plurality of existing content items of an audience having audience segments, wherein said content preferences indicate a performance metric for each content item and audience segment pair, wherein said plurality of existing content items comprise a plurality of content features indicating characteristics of a corresponding content item;

obtaining, for each of a plurality of advertisers, a weight for each of a plurality of said audience segments, wherein a given weight indicates a future interest of a given advertiser in a given audience segment;

calculating, using said at least one processing device, a pairwise similarity between a vector of said content preferences and a vector of said weights for said plurality of said audience segments to obtain a ranked list of said plurality of existing content items sorted by said pairwise similarity; and

generating, using said at least one processing device, a summarization of said content features, wherein said summarization of said content features is used to plan one or more new content items based on said ranked list, wherein said summarization of said content features is based on a frequency of the content features in the ranked list of said plurality of existing content items, wherein said one or more new content items are produced after said generating and comprise one or more content features from said summarization.

14. The system of claim 13 , wherein said performance metric comprises an average rating measured for a particular content item.

15. The system of claim 13 , wherein said weight is based on amount of money a given advertiser expects to spend on targeting each of said plurality of audience segments.

16. The system of claim 13 , further comprising the step of computing a normalization ratio to scale said vector of said weights for said plurality of said audience segments with said performance metrics.

17. The system of claim 13 , wherein said obtaining of said weight for said plurality of said audience segments considers one or more subsets of said plurality of advertisers with similar profiles.

18. The method of claim 7 , wherein said one or more subsets are defined by means of a clustering algorithm over a set of vectors of interest of each advertiser for said plurality of advertisers.

19. The computer program product of claim 12 , wherein said one or more subsets are defined by means of a clustering algorithm over a set of vectors of interest of each advertiser for said plurality of advertisers.

20. The system of claim 17 , wherein said one or more subsets are defined by means of a clustering algorithm over a set of vectors of interest of each advertiser for said plurality of advertisers.