Reducing processing requirements to correct for bias in ratings data having interdependencies among demographic statistics

Patent metadata

Field	Value
Publication number	US-10740774-B2
Application number	US-201514835401-A
Country	US
Kind code	B2
Filing date	Aug 25, 2015
Priority date	Jul 15, 2015
Publication date	Aug 11, 2020
Grant date	Aug 11, 2020

Abstract

Methods, apparatus, systems and articles of manufacture to correct for bias in ratings data are disclosed. Example methods disclosed herein include accessing information specifying a set of population strata having corresponding strata demographic statistics capable of being combined to determine a first set of baseline demographic statistics for a sample population that corresponds to a second set of baseline demographic statistics for a target population. Disclosed example methods also include determining a set of weights corresponding respectively to the set of population strata that reduces bias between the second set of baseline demographic statistics for the target population and the first set of baseline demographic statistics for the sample population when the set of weights is applied to the strata demographic statistics to determine the first set of baseline demographic statistics for the sample population. Disclosed example methods further include adjusting attributes associated with sample population individuals to determine ratings data.

First claim

What is claimed is:

1. A method to correct for bias in ratings data, the method comprising:
receiving beacon request messages from a plurality of devices via a network, the beacon request messages including respective media exposure data associated with corresponding ones of the plurality of devices;
transmitting, responsive to the beacon request messages, corresponding beacon response messages to the plurality of devices via the network, the beacon response messages to cause the plurality of devices to transmit subsequent beacon request messages to a first data source;
receiving, via the network, demographic data associated with the plurality of device from the first data source;
combining, by executing an instruction with a processor, the media exposure data and the demographic data to determine attributes associated with a sample population corresponding to the plurality of devices;
obtaining, via the network from a second data source different from the first data source and the plurality of devices, attributes associated with a target population;
accessing, by executing an instruction with the processor, information specifying a set of population strata having corresponding strata demographic statistics to be combined to determine a first set of baseline demographic statistics for the sample population that corresponds to a second set of baseline demographic statistics for the target population, the strata demographic statistics determined based on the attributes associated with the sample population, the second set of baseline demographic statistics determined based on the attributes associated with the target population;
forming, by executing an instruction with the processor, a matrix to be processed by the processor with a numerical processing technique, the matrix having matrix elements corresponding to pairings of ones of the strata demographic statistics and ones of the first set of baseline demographic statistics, a first matrix element corresponding to a pairing of a first one of the strata demographic statistics and a first one of the first set of baseline demographic statistics, the first matrix element determined by combining the first one of the strata demographic statistics with a first one of the second set of baseline demographic statistics based on a first value and a second value, the first one of the second set of baseline demographic statistics corresponding to the first one of the first set of baseline demographic statistics, the first value set to one or zero based on whether the first one of the strata demographic statistics is included in a numerator term used to calculate the first one of the first set of baseline demographic statistics based on the strata demographic statistics, the second value set to one or zero based on whether the first one of the strata demographic statistics is included in a denominator term used to calculate the first one of the first set of baseline demographic statistics;
determining, with the numerical processing technique performed with the processor, a set of weights based on the matrix, the set of weights corresponding respectively to the set of population strata, the set of weights to reduce bias between the second set of baseline demographic statistics for the target population and the first set of baseline demographic statistics for the sample population when the set of weights is applied to the strata demographic statistics to determine the first set of baseline demographic statistics for the sample population;
adjusting, with the processor based on the set of weights, the attributes associated with the sample population to determine the ratings data; and
transmitting the ratings data, via the network, to a receiving device different from the plurality of devices and the first and second data sources.

2. The method as defined in claim 1 , wherein the population strata form a mutually exclusive set that covers the sample population.

3. The method as defined in claim 1 , wherein the strata demographic statistics represent numbers of individuals from the sample population belonging to respective ones of the set of population strata.

4. The method as defined in claim 1 , wherein the determining of the set of weights based on the matrix includes determining the set of weights to minimize an expression based on multiplying the matrix and a vector of the weights subject to a constraint that the weights have values greater than or equal to one.

5. The method as defined in claim 1 , wherein a first individual of the sample population is included in a first one of the population strata, a second individual of the sample population is included in a second one of the population strata, and the adjusting of the attributes associated with the sample population to determine the ratings data includes:
scaling a contribution of an attribute associated with the first individual by a first one of the weights corresponding to the first one of the population strata to determine a first scaled attribute contribution;
scaling a contribution of an attribute associated with the second individual by a second one of the weights corresponding to the second one of the population strata to determine a second scaled attribute contribution; and
combining the first scaled attribute contribution and the second scaled attribute contribution to determine the ratings data.

6. The method as defined in claim 1 , wherein the transmitting of the ratings data includes transmitting the ratings data at least one of electronically or optically from the processor to the receiving device via the network.

7. A tangible computer readable storage medium comprising computer readable instructions which, when executed, cause a processor to at least:
receive beacon request messages from a plurality of devices via a network, the beacon request messages to include respective media exposure data associated with corresponding ones of the plurality of devices;
transmit, responsive to the beacon request messages, corresponding beacon response messages to the plurality of devices via the network, the beacon response messages to cause the plurality of devices to transmit subsequent beacon request messages to a first data source;
receive, via the network, demographic data associated with the plurality of device from the first data source;
combine the media exposure data and the demographic data to determine attributes associated with a sample population corresponding to the plurality of devices;
obtain, via the network from a second data source different from the first data source and the plurality of devices, attributes associated with a target population;
access information specifying a set of population strata having corresponding strata demographic statistics to be combined to determine a first set of baseline demographic statistics for the sample population that corresponds to a second set of baseline demographic statistics for the target population, the strata demographic statistics determined based on the attributes associated with the sample population, the second set of baseline demographic statistics determined based on the attributes associated with the target population;
form a matrix to be processed by the processor with a numerical processing technique, the matrix having matrix elements corresponding to pairings of ones of the strata demographic statistics and ones of the first set of baseline demographic statistics, a first matrix element corresponding to a pairing of a first one of the strata demographic statistics and a first one of the first set of baseline demographic statistics, the first matrix element to be determined by combining the first one of the strata demographic statistics with a first one of the second set of baseline demographic statistics

Assignees

• Nielsen Co Us Llc

Inventors

• Sheppard Michael
• Sullivan Jonathan
• Terrazas Alejandro
• Lipa Peter
• Perez Albert Ronald

Classifications

• G06Q30/0203

  Market surveys; Market polls · CPC title

• G06Q30/0204 Primary

  Market segmentation · CPC title

• G06Q30/0201

  Market modelling; Market analysis; Collecting market data · CPC title