Methods, systems, apparatus and articles of manufacture to determine causal effects
Methods, systems, apparatus and articles of manufacture to determine causal effects are disclosed herein. An example apparatus includes a weighting engine to calculate a first set of weights for a first set of covariates corresponding to a treatment dataset and a second set of weights for a second set of covariates corresponding to a control dataset using maximum entropy, the first set of weights to equal the second set of weights. The example apparatus also includes a weighting response engine to calculate a weighted response for the treatment dataset and a weighted response for the control dataset by: mapping the first set of weights and the second set of weights to a uniform weighting identifier, determining a constraint matrix based on the first set of weights, the second set of weights and the uniform weighting identifier, and bypassing multivariate reweighting by calculating the weighted response for the treatment dataset and the weighted response for the control dataset by applying maximum entropy to the constraint matrix.
1. An apparatus comprising:
a weighting engine to calculate a first set of weights for a first set of covariates corresponding to a treatment dataset and a second set of weights for a second set of covariates corresponding to a control dataset using maximum entropy, the first set of weights to equal the second set of weights;
a weighting response engine to calculate a weighted response for the treatment dataset and a weighted response for the control dataset by:
mapping the first set of weights and the second set of weights to a uniform weighting identifier;
determining a constraint matrix based on the first set of weights, the second set of weights and the uniform weighting identifier; and
bypassing multivariate reweighting to improve an operating efficiency of a computing device by calculating the weighted response for the treatment dataset and the weighted response for the control dataset by applying maximum entropy to the constraint matrix; and
a report generator to transmit a report to an audience measurement entity, the report generator to cause display of the report on a device, at least one of the weighting engine, the weighting response engine, or the report generator implemented by a logic circuit.
2. The apparatus of claim 1 , further including a covariate engine to determine the first set of covariates to be processed for the treatment dataset of an advertisement campaign and the second set of covariates to be processed for the control dataset of the advertisement campaign, the treatment dataset indicative of being exposed to an advertisement, the control dataset indicative of not being exposed to an advertisement.
3. The apparatus of claim 1 , wherein the weighting engine is to determine the first set of weights independent of the second set of weights to increase an operational efficiency of the apparatus.
4. The apparatus of claim 1 , wherein the weighting response engine is to calculate the weighted response for the treatment dataset and the weighted response for the control dataset so that the weighted response for the treatment dataset and the weighted response for the control dataset are on a common scale.
5. The apparatus of claim 1 , wherein the weighting response engine is to determine an average treatment effect by determining a difference between the weighted response for the treatment dataset and the weighted response for the control dataset, the average treatment effect indicative of a potential increase per individual exposed to an advertisement campaign.
6. The apparatus of claim 1 , wherein the report generator is to cause the display of the report via a webpage in a first state with a set of options, the options selectable by a user to change to the first state.
7. A non-transitory computer readable medium comprising instructions that, when executed, cause a machine to at least:
calculate a first set of weights for a first set of covariates corresponding to a treatment dataset and a second set of weights for a second set of covariates corresponding to a control dataset using maximum entropy, the first set of weights to equal the second set of weights; and
calculate a weighted response for the treatment dataset and a weighted response for the control dataset by:
mapping the first set of weights and the second set of weights to a uniform weighting identifier;
determining a constraint matrix based on the first set of weights, the second set of weights and the uniform weighting identifier;
bypassing multivariate reweighting to improve an operating efficiency of the machine by calculating the weighted response for the treatment dataset and the weighted response for the control dataset by applying maximum entropy to the constraint matrix; and
transmit a report to an audience measurement entity to cause display of the report on a device of the audience measurement entity.
8. The non-transitory computer readable medium of claim 7 , wherein the instructions further cause the machine to determine the first set of covariates to be processed for the treatment dataset of an advertisement campaign and the second set of covariates to be processed for the control dataset of the advertisement campaign, the treatment dataset indicative of being exposed to an advertisement, the control dataset indicative of not being exposed to an advertisement.
9. The non-transitory computer readable medium of claim 7 , wherein the instructions further cause the machine to determine the first set of weights independent of the second set of weights to increase an operational efficiency of the machine.
10. The non-transitory computer readable medium of claim 7 , wherein the instructions further cause the machine to calculate the weighted response for the treatment dataset and the weighted response for the control dataset so that the weighted response for the treatment dataset and the weighted response for the control dataset are on a common scale.
11. The non-transitory computer readable medium of claim 7 , wherein the instructions further cause the machine to determine an average treatment effect by determining a difference between the weighted response for the treatment dataset and the weighted response for the control dataset, the average treatment effect indicative of a potential increase per individual exposed to an advertisement campaign.
12. The non-transitory computer readable medium of claim 7 , wherein the instructions further cause the machine to display the report on the device via a webpage in a first state with a set of options, the options selectable by a user to change to the first state.
13. An apparatus comprising:
means for calculating a weight to calculate a first set of weights for a first set of covariates corresponding to a treatment dataset and a second set of weights for a second set of covariates corresponding to a control dataset using maximum entropy, the first set of weights to equal the second set of weights;
means for determining a weighted response to calculate a weighted response for the treatment dataset and a weighted response for the control dataset by:
mapping the first set of weights and the second set of weights to a uniform weighting identifier;
determining a constraint matrix based on the first set of weights, the second set of weights and the uniform weighting identifier; and
bypassing multivariate reweighting to improve an operating efficiency of a computing device by calculating the weighted response for the treatment dataset and the weighted response for the control dataset by applying maximum entropy to the constraint matrix; and
means for generating a report to display a report on a device of an audience measurement entity, at least one of the means for calculating a weight, the means for determining a weighted response, or the means for generating a report implemented by a logic circuit.
14. The apparatus of claim 13 , further including means for generating a covariate to determine the first set of covariates to be processed for the treatment dataset of an advertisement campaign and the second set of covariates to be processed for the control dataset of the advertisement campaign, the treatment dataset indicative of being exposed to an advertisement, the control dataset indicative of not being exposed to an advertisement.
15. The apparatus of claim 13 , wherein the weight calculating means is to determine the first set of weights independent of the second set of weights to increase an operational efficiency of the apparatus.
16. The apparatus of claim 13 , wherein the weighted response determining means is to calculate the weighted response for the treatment dataset and the weighted response for the control dataset so that the weighted response for the treatment dataset and the weighted response for the control dataset are on a common scale.
17. The apparatus of claim 13 , wherein the weighted response determining means is to determine an average treatment effect by determining a difference between the weighted response for the treatment dataset and the weighted response for the control dataset, the average treatment effect indicative of a potential increase per individual exposed to an advertisement campaign.
18. The apparatus of claim 13 , wherein the report generating means is to display the report on the device via a webpage in a first state with a set of options, the options selectable by a user to change to the first state.