IP classification
A set of Internet Protocol (IP) addresses is received wherein each IP address is associated with a corresponding set of features. For an IP address in the set, the IP address is evaluated based at least in part on a set of inclusion criteria. For the IP address in the set, a likelihood that the IP address is residential or non-residential is generated based at least in part on the corresponding set of features and the evaluation of the IP address based at least in part on the set of inclusion criteria. For the IP address in the set, a training sample is generated that includes the IP address, at least some of the corresponding set of features, and a label. A labeled training data set is output that includes the training sample, where an IP address classifier is trained using the labeled training data set.
1. A system, comprising:
a memory; and
a processor coupled to the memory and configured to:
receive a set of Internet Protocol (IP) addresses, wherein each IP address in the set of IP addresses is associated with a corresponding set of features;
for an IP address in the set of IP addresses, evaluate the IP address based at least in part on a set of inclusion criteria;
for the IP address in the set of IP addresses, generate a likelihood that the IP address is residential or non-residential based at least in part on the corresponding set of features and the evaluation of the IP address based at least in part on the set of inclusion criteria;
for the IP address in the set of IP addresses, generate a training sample that includes the IP address, at least some of the corresponding set of features, and a label; and
output a labeled training data set that includes the training sample, wherein an IP address classifier is trained using the labeled training data set.
2. The system of claim 1 , wherein:
the IP address classifier that is trained using the labeled training data set becomes a trained IP address classifier; and
the trained IP address classifier inputs a second set of unclassified IP addresses and outputs a second set of classified IP addresses.
3. The system of claim 2 , wherein:
a household group generator inputs the second set of classified IP addresses and generates a household group; and
content in a cache is managed based at least in part on the household group.
4. The system of claim 2 , wherein:
a household group generator inputs the second set of classified IP addresses and generates a household group; and
a configuration associated with a configurable network is adjusted based at least in part on one or more of the following: the household group or the second set of classified IP addresses.
5. The system of claim 1 , wherein the corresponding set of features includes one or more of the following: a number of broadband clients, a number of broadband sessions, a number of clients, a number of devices per day, a number of mobile devices, a number of non-stationary devices, a number of non-stationary clients, a number of publishers, or a Boolean value describing whether the IP address is a shared IP address.
6. The system of claim 1 , wherein the corresponding set of features includes one or more of the following: a median of a client count, a percentile of a client count, a percentile of a connected TV device model count, a median of a device model count, a percentile of a device model count, or a percentage of cellular sessions out of all sessions.
7. The system of claim 1 , wherein the corresponding set of features includes one or more of the following: session information or proprietary information.
8. The system of claim 1 , wherein the set of inclusion criteria includes one or more of the following: an IP address ratio, a special case, a proprietary-based blacklist, a super IP, or an is encountered bad case.
9. The system of claim 1 , wherein evaluating the IP address based at least in part on the set of inclusion criteria includes:
determining a number of same-account client identifier pairs;
determining a number of total client identifier pairs;
determining a ratio of the number of same-account client identifier pairs to the number of total client identifier pairs; and
comparing the ratio against a threshold.
10. The system of claim 1 , wherein evaluating the IP address based at least in part on the set of inclusion criteria includes:
determining a number of client identifiers associated with the IP address in the set of IP addresses; and
if the number of client identifiers equals one, excluding the IP address in the set of IP addresses.
11. The system of claim 1 , wherein evaluating the IP address based at least in part on the set of inclusion criteria includes:
determining a number of things that share the IP address in the set of IP addresses;
determining a number of broadband client identifiers associated with the IP address in the set of IP addresses;
determining a number of client identifiers associated with the IP address in the set of IP addresses; and
determining whether one or more threshold-based tests are exceeded using one or more of: the number of things, the number of broadband client identifiers, and the number of client identifiers.
12. The system of claim 1 , wherein evaluating the IP address based at least in part on the set of inclusion criteria includes:
determining a number of things that share the IP address in the set of IP addresses;
determining a proprietary-based blacklist associated with the IP address in the set of IP addresses;
determining a percentage of cellular sessions associated with the IP address in the set of IP addresses;
determining a number of devices associated with the IP address in the set of IP addresses; and
determining whether one or more threshold-based tests are exceeded using one or more of: the number of things, the proprietary-based blacklist, the percentage of cellular sessions, and the number of devices.
13. A method, comprising:
receiving a set of Internet Protocol (IP) addresses, wherein each IP address in the set of IP addresses is associated with a corresponding set of features;
for an IP address in the set of IP addresses, evaluating the IP address based at least in part on a set of inclusion criteria;
for the IP address in the set of IP addresses, generating a likelihood that the IP address is residential or non-residential based at least in part on the corresponding set of features and the evaluation of the IP address based at least in part on the set of inclusion criteria;
for the IP address in the set of IP addresses, generating a training sample that includes the IP address, at least some of the corresponding set of features, and a label; and
outputting a labeled training data set that includes the training sample, wherein an IP address classifier is trained using the labeled training data set.
14. The method of claim 13 , wherein:
the IP address classifier that is trained using the labeled training data set becomes a trained IP address classifier; and
the trained IP address classifier inputs a second set of unclassified IP addresses and outputs a second set of classified IP addresses.
15. The method of claim 14 , wherein:
a household group generator inputs the second set of classified IP addresses and generates a household group; and
content in a cache is managed based at least in part on the household group.
16. The method of claim 13 , wherein the corresponding set of features includes one or more of the following: a median of a client count, a percentile of a client count, a percentile of a connected TV device model count, a median of a device model count, a percentile of a device model count, or a percentage of cellular sessions out of all sessions.
17. The method of claim 13 , wherein the set of inclusion criteria includes one or more of the following: an IP address ratio, a special case, a proprietary-based blacklist, a super IP, or an encountered bad case.
18. The method of claim 13 , wherein evaluating the IP address based at least in part on the set of inclusion criteria includes:
determining a number of same-account client identifier pairs;
determining a number of total client identifier pairs;
determining a ratio of the number of same-account client identifier pairs to the number of total client identifier pairs; and
comparing the ratio against a threshold.
19. The method of claim 13 , wherein evaluating the IP address based at least in part on the set of inclusion criteria includes:
determining a number of client identifiers associated with the IP address in the set of IP addresses; and
if the number of client identifiers equals one, excluding the IP address in the set of IP addresses.
20. A computer program product embodied in a non-transitory computer readable medium and comprising computer instructions for:
receiving a set of Internet Protocol (IP) addresses, wherein each IP address in the set of IP addresses is associated with a corresponding set of features;
for an IP address in the set of IP addresses, evaluating the IP address based at least in part on a set of inclusion criteria;
for the IP address in the set of IP addresses, generating a likelihood that the IP address is residential or non-residential based at least in part on the corresponding set of features and the evaluation of the IP address based at least in part on the set of inclusion criteria;
for the IP address in the set of IP addresses, generating a training sample that includes the IP address, at least some of the corresponding set of features, and a label; and
outputting a labeled training data set that includes the training sample, wherein an IP is address classifier is trained using the labeled training data set.