Methods and apparatus to estimate demographics of a household

1. A method to determine demographics for non-panelist households, and reduce a quantity of panelist data retrieved from a computer memory, the method comprising:

reducing, by executing an instruction with a processor, the quantity of panelist data retrieved from the computer memory by calculating a first demographic constraint average and a second demographic constraint average in connection with tuning events of non-panelist households, the first and second demographic constraint averages based on (a) acquiring a first quantity of temporally-mapped panelist demographic distribution data occurring during and associated with a first tuning event of a set-top box of a respective one of the non-panelist households and (b) acquiring a second quantity of temporally-mapped panelist demographic distribution data occurring during and associated with a second tuning event of the set-top box of the respective one of the non-panelist households, the first and second tuning events of the non-panelist household reducing a quantity of the first and second temporally-mapped panelist demographic distribution data retrieved from the computer memory, the first quantity of temporally-mapped panelist subset of demographic distribution data including (a) a first probability that the first tuning event is associated with a first demographic constraint and (b) a second probability that the first tuning event is associated with a second demographic constraint, the second quantity of temporally-mapped panelist demographic distribution data including (a) a third probability that the second tuning event is associated with the first demographic constraint and (b) a fourth probability that the second tuning event is associated with the second demographic constraint;

based on the first demographic constraint average and the first probability that the first tuning event is associated with the first demographic constraint, determining, by executing an instruction with the processor, a first likelihood score of the non-panelist household being associated with the first demographic constraint;

based on the second demographic constraint average and the second probability that the first tuning event is associated with the second demographic constraint, determining, by executing an instruction with the processor, a second likelihood score of the non-panelist household being associated with the second demographic constraint;

estimating, by executing an instruction with the processor, a household characteristic of the non-panelist household based on the first likelihood score and the second likelihood score, the household characteristic including a number of members of the non-panelist household; and

producing, by executing an instruction with the processor, ratings of presented media for a mixed audience of non-panelist and panelist households based on the household characteristic.

2. The method as defined in claim 1 , further including estimating a number of televisions within respective ones of the non-panelist households and locations of the televisions within the respective ones of the non-panelist households.

3. The method as defined in claim 1 , further including estimating a demographic of a non-panelist household member.

4. The method as defined in claim 3 , wherein the estimating of the demographic of the non-panelist household member includes determining a marginal of a demographic dimension for the non-panelist household member, the demographic dimension including the first demographic constraint and the second demographic constraint.

5. The method as defined in claim 1 , further including:

dividing the first demographic constraint average of a first one of the non-panelist households by a first constraint probability of an average demographic distribution for consumption events of panelist households; and

dividing the second demographic constraint average of a second one of the non-panelist households by a second constraint probability of the average demographic distribution of panelist households.

6. The method as defined in claim 1 , wherein the estimating of the household characteristic of the non-panelist households is based on at least one of the first and second demographic constraint, a total number of minutes consumed by respective ones of the non-panelist households, or a number of minutes consumed by respective ones of the non-panelist households per predetermined time-period segments.

7. The method as defined in claim 1 , further including calculating a score vector based on the first likelihood score and the second likelihood score, the score vector to be applied to a machine learning classifier to estimate the household characteristic.

8. The method as defined in claim 7 , wherein the machine learning classifier is a decision tree ensemble.

9. The method as defined in claim 8 , further including:

training the decision tree ensemble on consumption data of a first panelist household; and

testing the decision tree ensemble on consumption data of a second panelist household, the second panelist household being different than the first panelist household.

10. The method as defined in claim 1 , further including identifying whether at least one of the first likelihood score or the second likelihood score is equal to or greater than a threshold value.

11. An apparatus to determine demographics for non-panelist households, and to reduce a quantity of panelist data retrieved from a computer memory, the apparatus comprising:

a score calculator to:

reduce the quantity of panelist data retrieved from the computer memory by calculating a first demographic constraint average and a second demographic constraint average in connection with tuning events of non-panelist households, the first and second demographic constraint averages based on (a) acquiring a first quantity of temporally-mapped panelist demographic distribution data occurring during and associated with a first tuning event of a set-top box of a respective one of the non-panelist households and (b) acquiring a second quantity of temporally-mapped panelist demographic distribution data occurring during and associated with a second tuning event of the set-top box of the respective one of the non-panelist households, the first and second tuning events of the non-panelist households reducing a quantity of the first and second temporally-mapped panelist demographic distribution data retrieved from the computer memory, the first quantity of temporally-mapped panelist demographic distribution data including (a) a first probability that the first tuning event is associated with a first demographic constraint and (b) a second probability that the first tuning event is associated with a second demographic constraint, the second quantity of temporally-mapped panelist demographic distribution data including (a) a third probability that the second tuning event is associated with the first demographic constraint and (b) a fourth probability that the second tuning event is associated with the second demographic constraint;

based on the first demographic constraint average and the first probability that the first tuning event is associated with the first demographic constraint, determine a first likelihood score of the non-panelist household being associated with the first demographic constraint;

based on the second demographic constraint average and the second probability that the second tuning event is associated with the second demographic constraint, determine a second likelihood score of the non-panelist household being associated with the second demographic constraint;

a household estimator to estimate a household characteristic of the non-panelist household based on the first likelihood score and the second likelihood score, the household characteristic including a number of members of the non-panelist household; and

produce ratings of presented media for a mixed audience of non-panelist and panelist households based on the household characteristic.

12. The apparatus as defined in claim 11 , wherein the household estimator is to estimate a number of televisions within respective ones of the non-panelist households and locations of the televisions within the respective ones of the non-panelist households.

13. The apparatus as defined in claim 11 , wherein the household estimator is to estimate a demographic of a non-panelist household member.

14. The apparatus as defined in claim 13 , wherein the household estimator is to determine a marginal of a demographic dimension of the non-panelist household member, the demographic dimension includes the first demographic constraint and the second demographic constraint.

15. The apparatus as defined in claim 11 , wherein the score calculator is to:

divide the first demographic constraint average of a first one of the non-panelist households by a first constraint probability of an average demographic distribution for consumption events of panelist households and;

divide the second demographic constraint average of a second one of the non-panelist households by a second constraint probability of the average demographic distribution of panelist households to determine the second likelihood.

16. The apparatus as defined in claim 11 , wherein the household estimator is to estimate the household characteristic of the non-panelist households based on at least one of the first and second demographic constraints, a total number of minutes consumed by respective ones of the non-panelist households, or a number of minutes consumed by respective ones of the non-panelist households per predetermined time-period segments.

17. The apparatus as defined in claim 11 , wherein the score calculator is to calculate a score vector based on the first likelihood and the second likelihood, and the household estimator is to apply the score vector to a machine learning classifier to estimate the household characteristic.

18. The apparatus as defined in claim 17 , wherein the machine learning classifier is a decision tree ensemble.

19. The apparatus as defined in claim 18 , further including a decision tree trainer to:

train the decision tree ensemble on consumption data of a first panelist household; and

test the decision tree ensemble on consumption data of a second panelist household, the second panelist household being different than the first panelist household.

20. The apparatus as defined in claim 11 , wherein the household estimator is to identify whether at least one of the first likelihood or the second likelihood is equal to or greater than a threshold value.

21. A tangible computer readable storage medium to determine demographics for non-panelist households, and a quantity of panelist data retrieved from a computer memory, the tangible computer readable storage medium comprising instructions which, when executed, cause a machine to at least:

reduce the quantity of panelist data retrieved from the computer memory by calculating a first demographic constraint average and a second demographic constraint average in connection with tuning events of non-panelist households, the first and second demographic constraint averages based on (a) acquiring a first quantity of temporally-mapped panelist demographic distribution data occurring during and associated with a first tuning event of a set-top box of a respective one of the non-panelist households and (b) acquiring a second quantity of temporally-mapped panelist demographic distribution data occurring during and associated with a second tuning event of the set-top box of the respective one of the non-panelist households, the first and second tuning events of the non-panelist households reducing the quantity of the first and second temporally-mapped panelist demographic distribution data retrieved from the computer memory, the first quantity of temporally-mapped panelist demographic distribution data including (a) a first probability that the first tuning event is associated with a first demographic constraint and (b) a second probability that the first tuning event is associated with a second demographic constraint, the second quantity of temporally-mapped panelist demographic distribution data including (a) a third probability that the second tuning event is associated with the first demographic constraint and (b) a fourth probability that the second tuning event is associated with the second demographic constraint;

based on the first constraint average and the first probability that the first tuning event is associated with the first demographic constraint, determine a first likelihood score of the non-panelist household being associated with the first demographic constraint;

based on the second constraint average and the second probability that the first tuning event is associated with the second demographic constraint, determine a second likelihood score of the non-panelist household being associated with the second demographic constraint;

estimate a household characteristic of the non-panelist household based on the first likelihood score and the second likelihood score, the household characteristic including a number of members of the non-panelist household; and

produce ratings of presented media for a mixed audience of non-panelist and panelist households based on the household characteristic.

22. The tangible computer readable storage medium as defined in claim 21 , wherein the instructions, when executed, further cause the machine to estimate a number of televisions within respective ones of the non-panelist households and locations of the televisions within the respective ones of the non-panelist households.

23. The tangible computer readable storage medium as defined in claim 21 , wherein the instructions, when executed, further cause the machine to estimate a demographic of a non-panelist household member.

24. The tangible computer readable storage medium as defined in claim 23 , wherein the instructions, when executed, further cause the machine to determine a marginal of a demographic dimension for the non-panelist household member, the demographic dimension includes the first demographic constraint and the second demographic constraint.

25. The tangible computer readable storage medium as defined in claim 21 , wherein the instructions, when executed, further cause the machine to:

divide the first demographic constraint average of a first one of the non-panelist households by a first constraint probability of an average demographic distribution for consumption events of panelist households; and

divide the second demographic constraint average of a second one of the non-panelist households by a second constraint probability of the average demographic distribution of the panelist households to determine the second likelihood.

26. The tangible computer readable storage medium as defined in claim 21 , wherein the instructions, when executed, further cause the machine to estimate the household characteristic of the non-panelist households based on at least one of the first and second demographic constraints, a total number of minutes consumed by respective ones of the non-panelist households, or a number of minutes consumed by respective ones of the non-panelist households per predetermined time-period segments.

27. The tangible computer readable storage medium as defined in claim 21 , wherein the instructions, when executed, further cause the machine to calculate a score vector based on the first likelihood and the second likelihood, the score vector to be applied to a machine learning classifier to estimate the household characteristic.

28. The tangible computer readable storage medium as defined in claim 27 , wherein the machine learning classifier is a decision tree ensemble.

29. The tangible computer readable storage medium as defined in claim 28 , wherein the instructions, when executed, further cause the machine to:

train a decision tree ensemble on consumption data of a first panelist household; and

test the decision tree ensemble on consumption data of a second panelist household, the second panelist household being different than the first panelist household.

30. The tangible computer readable storage medium as defined in claim 21 , wherein the instructions, when executed, further cause the machine to identify whether at least one of the first likelihood or the second likelihood is equal to or greater than a threshold value.