Granted Patent B2

US 11,004,094 · App. 15/375,719 · Granted May 11, 2021

Abstract

Claims (71)

1. A computerized method for substantially eliminating bias from survey data while excluding personal survey information, the method comprising:

partitioning, with one or more processors, a reference data set into a plurality of reference data partitions using a data partitioning scheme, the reference data set comprising behavior data of a target population, each reference data partition associated with a characteristic;

partitioning, with the one or more processors, a subject data set into a plurality of subject data partitions using the data partitioning scheme, the subject data set comprising survey data including behavior data of a participant population, each subject data partition associated with a characteristic that corresponds to the characteristic associated with a reference data partition of the plurality of reference data partitions, wherein partitioning the subject data set includes excluding personal information from the subject data set;

identifying, with the one or more processors, a variable that is not present in the subject data set;

calculating, with the one or more processors, a value of the variable for each reference data partition based on a rate of occurrence of the variable in each reference data partition;

applying, with the one or more processors, the calculated value of the variable for each reference data partition to the subject data partition associated with the characteristic that corresponds to the characteristic associated with each reference data partition;

calculating, with the one or more processors, a statistic for each subject data partition after applying the calculated value of the variable for each reference data partition to the subject data partition associated with the characteristic that corresponds to the characteristic associated with each reference data partition;

calculating, with the one or more processors, a weight for each subject data partition based on a distribution of the target population among the plurality of reference data partitions;

generating, with the one or more processors, an adjusted statistic by applying the calculated weight for each subject data partition to the calculated statistic for each subject data partition to be more representative of the target population; and

displaying, on a user interface, the statistic to a user, wherein the statistic does not reflect bias from the survey data and excludes personal information from the subject data set.

2. The computerized method of claim 1 , wherein each reference data partition comprises users having the characteristic associated with the reference data partition,

wherein each subject data partition comprises users having the characteristic associated with the subject data partition,

wherein calculating the weight for each subject data partition comprises:

determining a reference portion for each reference data partition by dividing a number of the users in each reference data partition by a total number of users in the reference data set;

determining a subject portion for each subject data partition by dividing a number of the users in each subject data partition by a total number of users in the subject data set; and

dividing the subject portion for each subject data partition by the reference portion for the reference data partition associated with the characteristic to which the characteristic associated with each subject data partition corresponds.

3. The computerized method of claim 1 , further comprising reducing, with the one or more processors, a bias in the reference data set by adjusting the reference data set to comprise behavior data of users in the target population based on a plurality of users represented by desired population census data.

4. The computerized method of claim 3 , wherein partitioning the reference data set into the plurality of reference data partitions is performed after reducing the bias in the reference data set.

5. The computerized method of claim 1 , wherein the calculated value of the variable comprises one of: a mean value, an interval, and a vector of probabilities for each response associated with the variable.

6. The computerized method of claim 1 , wherein the data partitioning scheme is based on one of: television viewing behavior of users, and web browsing behavior of users.

7. The computerized method of claim 6 , wherein the statistic is a number of users that viewed specific content.

8. The computerized method of claim 1 , wherein the data partitioning scheme is based on mobile device usage behavior of users.

9. The computerized method of claim 1 , further comprising:

generating, with the one or more processors, a modeled variable that is not present in the reference data set and the subject data set;

calculating, with the one or more processors, a value of the modeled variable based on a rate of occurrence of the modeled variable in an external data set; and

applying, with the one or more processors, the calculated value of the modeled variable to one or more of a reference data partition of the plurality of reference data partitions and a subject data partition of the plurality of subject data partitions.

10. A non-transitory computer-readable medium encoded with instructions that, when executed by one or more processors, cause the one or more processors to perform a computerized method for substantially eliminating bias from survey data while excluding personal survey information, the method comprising:

partitioning a reference data set into a plurality of reference data partitions using a data partitioning scheme, the reference data set comprising behavior data of a target population, each reference data partition associated with a characteristic;

partitioning a subject data set into a plurality of subject data partitions using the data partitioning scheme, the subject data set comprising survey data including behavior data of a participant population, each subject data partition associated with a characteristic that corresponds to the characteristic associated with a reference data partition of the plurality of reference data partitions, wherein partitioning the subject data set includes excluding personal information from the subject data set;

identifying a variable that is not present in the subject data set;

calculating a value of the variable for each reference data partition based on a rate of occurrence of the variable in each reference data partition;

applying the calculated value of the variable for each reference data partition to the subject data partition associated with the characteristic that corresponds to the characteristic associated with each reference data partition;

calculating a statistic for each subject data partition after applying the calculated value of the variable for each reference data partition to the subject data partition associated with the characteristic that corresponds to the characteristic associated with each reference data partition;

calculating a weight for each subject data partition based on a distribution of the target population among the plurality of reference data partitions;

generating an adjusted statistic by applying the calculated weight for each subject data partition to the calculated statistic for each subject data partition to be more representative of the target population; and

displaying the statistic to a user on a user interface, wherein the statistic does not reflect bias from the survey data and excludes personal information from the subject data set.

11. The computerized method of claim 10 , further comprising:

calculating, with the one or more processors, a value of the modeled variable for each external data partition based on a rate of occurrence of the modeled variable in each external data partition; and

applying, with the one or more processors, the calculated value of the modeled variable for each external data partition to the reference data partition or subject data partition associated with the characteristic to which the characteristic associated with each external data partition corresponds.

12. A non-transitory computer-readable medium encoded with instructions that, when executed by one or more processors, cause the one or more processors to perform a computerized method for substantially eliminating bias from survey data, the method comprising:

partitioning a reference data set into a plurality of reference data partitions using a data partitioning scheme, the reference data set comprising behavior data of a target population, each reference data partition associated with a characteristic;

partitioning a subject data set into a plurality of subject data partitions using the data partitioning scheme, the subject data set comprising behavior data of a participant population, each subject data partition associated with a characteristic that corresponds to the characteristic associated with a reference data partition of the plurality of reference data partitions, wherein partitioning the subject data set includes excluding personal information from the subject data set;

identifying a variable that is not present in the subject data set;

calculating a value of the variable for each reference data partition based on a rate of occurrence of the variable in each reference data partition;

applying the calculated value of the variable for each reference data partition to the subject data partition associated with the characteristic that corresponds to the characteristic associated with each reference data partition;

calculating a statistic for each subject data partition after applying the calculated value of the variable for each reference data partition to the subject data partition associated with the characteristic that corresponds to the characteristic associated with each reference data partition;

calculating a weight for each subject data partition based on a distribution of the target population among the plurality of reference data partitions;

generating an adjusted statistic by applying the calculated weight for each subject data partition to the calculated statistic for each subject data partition to be more representative of the target population; and

displaying the statistic to a user on a user interface.

13. The non-transitory computer-readable medium of claim 12 , wherein the calculated value of the variable comprises one of: a mean value, an interval, and a vector of probabilities for each response associated with the variable.

14. The non-transitory computer-readable medium of claim 12 , wherein the method further comprises:

generating a modeled variable that is not present in the reference data set and the subject data set;

calculating a value of the modeled variable based on a rate of occurrence of the modeled variable in an external data set; and

applying the calculated value of the modeled variable to one or more of a reference data partition of the plurality of reference data partitions and a subject data partition of the plurality of subject data partitions.

15. A system, comprising:

one or more processors;

a memory storing instructions that, when executed, cause the one or more processors to perform a computerized method for substantially eliminating bias from survey data while excluding personal survey information, the method comprising:

partitioning a reference data set into a plurality of reference data partitions using a data partitioning scheme, the reference data set comprising behavior data of a target population, each reference data partition associated with a characteristic;

partitioning a subject data set into a plurality of subject data partitions using the data partitioning scheme, the subject data set comprising survey data including behavior data of a participant population, each subject data partition associated with a characteristic that corresponds to the characteristic associated with a reference data partition of the plurality of reference data partitions, wherein partitioning the subject data set includes excluding personal information from the subject data set;

identifying a variable that is not present in the subject data set;

calculating a value of the variable for each reference data partition based on a rate of occurrence of the variable in each reference data partition;

applying the calculated value of the variable for each reference data partition to the subject data partition associated with the characteristic that corresponds to the characteristic associated with each reference data partition;

calculating a statistic for each subject data partition after applying the calculated value of the variable for each reference data partition to the subject data partition associated with the characteristic that corresponds to the characteristic associated with each reference data partition;

calculating a weight for each subject data partition based on a distribution of the target population among the plurality of reference data partitions;

generating an adjusted statistic by applying the calculated weight for each subject data partition to the calculated statistic for each subject data partition to be more representative of the target population; and

displaying the statistic to a user on a user interface, wherein the statistic does not reflect bias from the survey data and excludes personal information from the subject data set.

16. The system of claim 15 , wherein the calculated value of the variable comprises one of: a mean value, an interval, and a vector of probabilities for each response associated with the variable.

17. The system of claim 15 , wherein the method further comprises:

generating a modeled variable that is not present in the reference data set and the subject data set;

calculating a value of the modeled variable based on a rate of occurrence of the modeled variable in an external data set; and

applying the calculated value of the modeled variable to one or more of a reference data partition of the plurality of reference data partitions and a subject data partition of the plurality of subject data partitions.

Assignments (9)