MEDIA DEVICE HOUSEHOLDING AND DEDUPLICATION

1 . An apparatus comprising:

at least one memory;

instructions; and

processor circuitry to execute the instructions to:

deduplicate at least one of panel data, automatic content recognition data, or return path data to generate a deduplicated data set;

generate a graph that connects devices identified in the deduplicated data set, the devices connected based on at least one of a shared internet protocol addresses, a shared location identifier, or a common device identifier assigned by a common device identification algorithm; and

remove one or more connections based on an inconsistency between the one or more connections and personally identifiable information.

2 . The apparatus of claim 1 , wherein the processor circuitry is to identify a segment of the graph including devices identified in the deduplicated data set that are associated with a single household.

3 . The apparatus of claim 2 , wherein the processor circuitry is to identify the segment based on a connected component analysis that assigns quality values to connections of the graph.

4 . The apparatus of claim 3 , wherein the processor circuitry is to identify connections of the graph with quality values greater than a threshold value as intra-household connections, and wherein quality values less than the threshold value are identified as inter-household connections.

5 . The apparatus of claim 1 , wherein the at least one of panel data, the automatic content recognition data, or the return path data is obtained from at least two different providers.

6 . The apparatus of claim 1 , wherein the processor circuitry is to remove a first one of the devices from the graph that is identified as inactive.

7 . The apparatus of claim 6 , wherein the processor circuitry is to reconstruct the graph after the first one of the devices is removed from the graph.

8 . A non-transitory machine readable storage medium comprising instructions which, when executed by processor circuitry, cause the processor circuitry to:

deduplicate at least one of panel data, automatic content recognition data, or return path data to generate a deduplicated data set;

generate a graph that connects devices identified in the deduplicated data set, the devices connected based on at least one of a shared internet protocol addresses, a shared location identifier, or a common device identifier assigned by a common device identification algorithm; and

remove one or more connections based on an inconsistency between the one or more connections and personally identifiable information.

9 . The non-transitory machine readable medium of claim 8 , wherein the instructions, when executed, cause the processor circuitry to identify a segment of the graph including devices identified in the deduplicated data set that are associated with a single household.

10 . The non-transitory machine readable medium of claim 9 , wherein the instructions, when executed, cause the processor circuitry to identify the segment based on a connected component analysis that assigns quality values to connections of the graph.

11 . The non-transitory machine readable medium of claim 10 , wherein the instructions, when executed, cause the processor circuitry to identify connections of the graph with quality values greater than a threshold value as intra-household connections, and wherein quality values less than the threshold value are identified as inter-household connections.

12 . The non-transitory machine readable medium of claim 8 , wherein the at least one of panel data, the automatic content recognition data, or the return path data is obtained from at least two different providers.

13 . The non-transitory machine readable medium of claim 8 , wherein the instructions, when executed, cause the processor circuitry to remove a first one of the devices from the graph that is identified as inactive.

14 . The non-transitory machine readable medium of claim 13 , wherein the instructions, when executed, cause the processor circuitry to reconstruct the graph after the first one of the devices is removed from the graph.

15 . A method comprising:

deduplicating, by executing an instruction with processor circuitry, at least one of panel data, automatic content recognition data, or return path data to generate a deduplicated data set;

generating, by executing an instruction with the processor circuitry, a graph that connects devices identified in the deduplicated data set, the devices connected based on at least one of a shared internet protocol addresses, a shared location identifier, or a common device identifier assigned by a common device identification algorithm; and

removing, by executing an instruction with processor circuitry, one or more connections based on an inconsistency between the one or more connections and personally identifiable information.

16 . The method of claim 15 , further including identifying a segment of the graph including devices identified in the deduplicated data set that are associated with a single household.

17 . The method of claim 16 , further including identifying the segment based on a connected component analysis that assigns quality values to connections of the graph.

18 . The method of claim 17 , further including identifying connections of the graph with quality values greater than a threshold value as intra-household connections, and wherein the quality values less than the threshold value are identified as inter-household connections.

19 . The method of claim 15 , wherein the at least one of panel data, the automatic content recognition data, or the return path data is obtained from at least two different providers.

20 . The method of claim 15 , further including removing a first one of the devices from the graph that is identified as inactive.

21 - 25 . (canceled)