Methods and apparatus to compensate for server-generated errors in database proprietor impression data due to misattribution and/or non-coverage

Patent metadata

Field	Value
Publication number	US-12045845-B2
Application number	US-202318103276-A
Country	US
Kind code	B2
Filing date	Jan 30, 2023
Priority date	Mar 13, 2014
Publication date	Jul 23, 2024
Grant date	Jul 23, 2024

Abstract

Methods and apparatus to compensate impression data for misattribution and non-coverage are disclosed. An example method includes receiving a first request from a first type of computing device; sending a request for demographic information corresponding to requests received at a first internet domain from the first type of computing device; generating an aggregated audience distribution including a first audience distribution of a first household aggregated with a second audience distribution of a second household; normalizing the aggregated audience distribution to generate a misattribution correction matrix, the misattribution correction matrix including a probability that an impression of the media is attributable to a first demographic group when the database proprietor determines the impression corresponds to a person in a second demographic group; and compensating misattribution error in the impressions by re-assigning the impressions from the second demographic group to the first demographic group using the misattribution correction matrix.

First claim

What is claimed is:

1. An audience measurement computing system for predicting genre corresponding to media impressions that are not encoded with genre information, the computing system comprising:
a network interface;
a processor;
memory having stored thereon machine-readable instructions that, when executed by the processor cause performance of operations comprising:
receiving, via the network interface and a first network connection with computing devices located at a plurality of households, first network communications indicative of access to media;
transmitting, to a database proprietor server located remotely from the computing system, and via the network interface and a second network connection with the database proprietor server, second network communications indicative of requests for demographic information corresponding to non-genre media impressions logged based on the first network communications;
receiving, from the database proprietor server, and via the network interface and the second network connection, third network communications indicative of demographic information corresponding to at least a portion of the non-genre media impressions;
based on the first network communications and the third network communications, logging non-genre impression records;
transforming the non-genre impression records into categorized impression data based on information included in the non-genre impression records;
applying a genre prediction model to the categorized impression data to generate genre predictions for the non-genre impression records, the genre prediction model including a classification rate indicative of an accuracy of the prediction model;
generating predicted-genre impressions by associating the genre predictions with the non-genre impression records, wherein the predicted-genre impressions comprise (i) a computer-generated misattribution error in which a first predicted-genre impression of the predicted-genre impressions is misattributed to a first demographic group, different from a second demographic group of a first user that accessed the media to which the first predicted-genre impression corresponds and (ii) a computer-generated non-coverage error in which the database proprietor server was unable to access a demographic group of a second user that accessed the media to which a second predicted-genre impression corresponds;
correcting the computer-generated misattribution error by reassigning the first predicted-genre impression to the second demographic group using a misattribution matrix; and
correcting the computer-generated non-coverage error by (i) determining the demographic group of the second user based on a non-coverage factor and (ii) assigning the second predicted-genre impression to the determined demographic group.

2. The computing system of claim 1 , the operations further comprising:
selecting variables in ones of the non-genre impression records in training data.

3. The computing system of claim 2 , wherein selecting the variables comprises selecting the variables by performing statistical significance testing on the variables.

4. The computing system of claim 2 , the operations further comprising:
excluding at least some of the ones of the non-genre impression records from the training data based on a selection criterion.

5. The computing system of claim 2 , the operations further comprising:
constructing a candidate model using the variables.

6. The computing system of claim 5 , the operations further comprising:
evaluating the candidate model based on test data from a historical database; and
calculating a second classification rate of the candidate model.

7. The computing system of claim 6 , the operations further comprising:
adjusting the variables after a determination that the second classification rate does not satisfy a threshold.

8. The computing system of claim 6 , the operations further comprising:
comparing the second classification rate of the candidate model to third classification rates of second candidate models to determine which of the candidate model or the second candidate models corresponds to a highest accuracy.

9. A non-transitory computer readable storage medium having stored thereon machine-readable instructions that, when executed by a processor of an audience measurement computing system, cause performance of operations comprising:
receiving, via a network interface and a first network connection with computing devices located at a plurality of households, first network communications indicative of access to media;
transmitting, to a database proprietor server located remotely from the computing system, and via the network interface and a second network connection with the database proprietor server, second network communications indicative of requests for demographic information corresponding to non-genre media impressions logged based on the first network communications;
receiving, from the database proprietor server, and via the network interface and the second network connection, third network communications indicative of demographic information corresponding to at least a portion of the non-genre media impressions;
based on the first network communications and the third network communications, logging non-genre impression records;
transforming the non-genre impression records into categorized impression data based on information included in the non-genre impression records;
applying a genre prediction model to the categorized impression data to generate genre predictions for the non-genre impression records, the genre prediction model including a classification rate indicative of an accuracy of the prediction model; and
generating predicted-genre impressions by associating the genre predictions with the non-genre impression records, wherein the predicted-genre impressions comprise (i) a computer-generated misattribution error in which a first predicted-genre impression of the predicted-genre impressions is misattributed to a first demographic group, different from a second demographic group of a first user that accessed the media to which the first predicted-genre impression corresponds and (ii) a computer-generated non-coverage error in which the database proprietor server was unable to access a demographic group of a second user that accessed the media to which a second predicted-genre impression corresponds;
correcting the computer-generated misattributing error by reassigning the first predicted-genre impression to the second demographic group using a misattribution matrix; and
correcting the computer-generated non-coverage error by (i) determining the demographic group of the second user based on a non-coverage factor and (ii) assigning the second predicted-genre impression to the determined demographic group.

10. The non-transitory computer readable storage medium of claim 9 , the operations further comprising:
selecting variables in ones of the non-genre impression records in training data.

11. The non-transitory computer readable storage medium of claim 10 , wherein selecting the variables comprises selecting the variables by performing statistical significance testing on the variables.

12. The non-transitory computer readable storage medium of claim 10 , the operations further comprising:
excluding at least some of the ones of the non-genre impression records from the training data based on a selection criterion.

13. The non-transitory computer readable storage medium of claim 10 , the operations further comprising:
constructing a candidate model using the variables.

14. The non-transitory computer readable stor

Assignees

• Nielsen Co Us Llc

Inventors

• Rao Kumar Nagaraja
• Luo Tianjue
• Perez Albert Ronald
• Bell Stephen S
• Zhang Mimi
• Wong David
• Haskell Jennifer

Classifications

• H04W4/21

  for social networking applications · CPC title

• H04W4/18

  Information format or content conversion, e.g. adaptation by the network of the transmitted or received information for the purpose of wireless delivery to users or terminals · CPC title

• G06Q30/0242

  Determining effectiveness of advertisements · CPC title

• G06Q30/0241

  Advertisements · CPC title

• G06Q30/0245

  Surveys · CPC title