System and method for approximating numerical features via cubic splines and applications thereof
AssigneeYahoo Ad Tech LLC
InventorAlex Shtoff
Life of this patent
Estimating the term…
Abstract
The present teaching relates to method, system, medium, and implementations for approximating a non-linear relationship between a numerical feature and an output of a model. A value of a numerical feature is received and is transformed, via a transform function, into a transformed value within a fixed range. With respect to each of a plurality of basis functions used for approximating the non-linear relationship, a respective basis function value of the basis function is computed based on the transformed value. An approximated value of the non-linear numeric feature is generated based on a sum of the plurality of basis function values weighted respectively by each corresponding one of a set of the weights, obtained via machine learning.
Claims20
A method of data representation in online advertising, the method comprising:
- collecting training data from online events, wherein the training data comprises information related to users' interactions with an advertisement from an advertiser on a website;
- generating, based on the training data, a feature vector comprising:
- an involvement feature including a plurality of base functions of a value of involvement, wherein each of the plurality of base functions is applied with a corresponding machine-learned weight, and a weighted summation of the plurality of base functions of the value of involvement is an approximation of the value of involvement,
- a demographic feature indicative of demographic information of each of the users, and
- an advertiser feature indicative of the advertiser;
- and
- providing, based on the feature vector, a determination in terms of which advertisement to be displayed on a given website to a user having known demographic information.
Machine readable and non-transitory medium having information recorded thereon for data representation in online advertising, wherein the information, once read by the machine, causes the machine to perform the following steps:
- collecting training data from online events, wherein the training data comprises information related to users' interactions with an advertisement from an advertiser on a website;
- generating, based on the training data, a feature vector comprising:
- an involvement feature including a plurality of base functions of a value of involvement, wherein each of the plurality of base functions is applied with a corresponding machine-learned weight, and a weighted summation of the plurality of base functions of the value of involvement is an approximation of the value of involvement,
- a demographic feature indicative of demographic information of each of the users, and
- an advertiser feature indicative of the advertiser;
- and
- providing, based on the feature vector, a determination in terms of which advertisement to be displayed on a given website to a user having known demographic information.
A system for data representation in online advertising, the system comprising:
- memory storing computer program instructions;
- and
- one or more processors that, in response to executing the computer program instructions, effectuate operations comprising:
- collecting training data from online events, wherein the training data comprises information related to users' interactions with an advertisement from an advertiser on a website;
- generating, based on the training data, a feature vector comprising:
- an involvement feature including a plurality of base functions of a value of involvement, wherein each of the plurality of base functions is applied with a corresponding machine-learned weight, and a weighted summation of the plurality of base functions of the value of involvement is an approximation of the value of involvement,
- a demographic feature indicative of demographic information of each of the users, and
- an advertiser feature indicative of the advertiser;
- and
- providing, based on the feature vector, a determination in terms of which advertisement to be displayed on a given website to a user having known demographic information.
Dependencies are inferred from the claim text.