Back to results
US 9,319,486 B2Granted
Predicting interest levels associated with publication and content item combinations
AssigneeGoogle Inc.
InventorsLuciano Sbaiz, Dimitre Trendafilov
Life of this patent
Estimating the term…
Abstract
A method of predicting interest levels associated with publication and content item combinations is described. Additionally, a server computing device for predicting interest levels associated with publication and content item combinations is described. Further, a computer-readable storage device having processor-executable instructions embodied thereon is described. The processor-executable instructions are for predicting interest levels associated with publication and content item combinations.
Claims23
A method of predicting interest levels associated with publication and content item combinations, the method implemented by a server computing device communicatively coupled to at least one client computing device, the method comprising:
- determining a first interest level associated with a first combination including a first publication and a first set of keywords used for selecting a first content item for display in conjunction with the first publication, the first interest level representing a click-through-rate associated with the first content item and the first publication;
- determining a second interest level associated with a second combination including a second publication and a second set of keywords used for selecting a second content item for display in conjunction with the second publication, the second interest level representing a click-through-rate associated with the second content item and the second publication;
- identifying a first dependency that indicates a proportion of the first interest level attributable to the first set of keywords;
- identifying a second dependency that indicates a proportion of the first interest level attributable to a first provider of publications associated with the first publication;
- identifying a third dependency that indicates a proportion of the second interest level attributable to the second set of keywords;
- identifying a fourth dependency that indicates a proportion of the second interest level attributable to a second provider of publications associated with the second publication;
- and
- generating (i) a proposed combination of the first publication and one of the first set of keywords and the second set of keywords and (ii) a predicted interest level associated with the proposed combination, based on the first interest level, the second interest level, the first dependency, the second dependency, the third dependency, and the fourth dependency, wherein generating the predicted interest level comprises determining a predicted interest level associated with each combination of the first publication and one of the first set of keywords and the second set of keywords, and generating the proposed combination comprises identifying which combination of the first publication and one of the first set of keywords and the second set of keywords has a highest predicted interest level.
5 more elements
A server computing device for predicting interest levels associated with publication and content item combinations, said server computing device is communicatively coupled to a database and to at least one client computing device, said server computing device is configured to:
- determine a first interest level associated with a first combination including a first publication and a first set of keywords used for selecting a first content item for display in conjunction with the first publication, the first interest level representing a click-through-rate associated with the first content item when displayed with the first publication;
- determine a second interest level associated with a second combination including a second publication and a second set of keywords used for selecting a second content item for display in conjunction with the second publication, the second interest level representing a click-through-rate associated with the second content item when displayed with the second publication;
- identify a first dependency that indicates a proportion of the first interest level attributable to the first set of keywords;
- identify a second dependency that indicates a proportion of the first interest level attributable to a first provider of publications associated with the first publication;
- identify a third dependency that indicates a proportion of the second interest level attributable to the second set of keywords;
- identify a fourth dependency that indicates a proportion of the second interest level attributable to a second provider of publications associated with the second publication;
- and
- generate (i) a proposed combination of the first publication and one of the first set of keywords and the second set of keywords and (ii) a predicted interest level associated with the proposed combination, based on the first interest level, the second interest level, first dependency, the second dependency, the third dependency, and the fourth dependency, wherein generating the predicted interest level comprises determining a predicted interest level associated with each combination of the first publication and one of the first set of keywords and the second set of keywords, and generating the proposed combination comprises identifying which combination of the first publication and one of the first set of keywords and the second set of keywords has a highest predicted interest level.
5 more elements
A non-transitory computer-readable storage device having processor-executable instructions embodied thereon, for predicting interest levels associated with publication and content item combinations, wherein when executed by a server computing device communicatively coupled to a database and to the at least one client computing device, the processor-executable instructions cause the server computing device to:
- determine a first interest level associated with a first combination including a first publication and a first set of keywords used for selecting a first content item for display in conjunction with the first publication, the first interest level representing a click-through-rate associated with the first content item when displayed with the first publication;
- determine a second interest level associated with a second combination including a second publication and a second set of keywords used for selecting a second content item for display in conjunction with the second publication, the second interest level representing a click-through-rate associated with the second content item when displayed with the second publication;
- identify a first dependency that indicates a proportion of the first interest level attributable to the first set of keywords;
- identify a second dependency that indicates a proportion of the first interest level attributable to a first provider of publications associated with the first publication;
- identify a third dependency that indicates a proportion of the second interest level attributable to the second set of keywords;
- identify a fourth dependency that indicates a proportion of the second interest level attributable to a second provider of publications associated with the second publication;
- and
- generate (i) a proposed combination of the first publication and one of the first set of keywords and the second set of keywords and (ii) a predicted interest level associated with the proposed combination, based on the first interest level, the second interest level, first dependency, the second dependency, the third dependency, and the fourth dependency, wherein generating the predicted interest level comprises determining a predicted interest level associated with each combination of the first publication and one of the first set of keywords and the second set of keywords, and generating the proposed combination comprises identifying which combination of the first publication and one of the first set of keywords and the second set of keywords has a highest predicted interest level.
5 more elements
Dependencies are inferred from the claim text.