System and method for demographic profiling of mobile terminal users based on network-centric estimation of installed mobile applications and their usage patterns
Methods and systems for creating demographic profiles of mobile communication network users. A demographic classification system analyzes network traffic, so as to estimate the specific combination of application classes installed on a given terminal, and usage patterns of the applications over time. This combination of application classes and their respective usage patterns are a highly personalized choice made by the user, and is therefore used by the system to deduce the user's demographic profile. The demographic classification system operates on monitored network traffic, as opposed to obtaining explicit and accurate information regarding the installed applications from the terminal. The system then deduces the demographic profile of the user from the list of estimated application classes.
1. A system, comprising:
an interface, which is configured to receive traffic from a mobile network; and
a processor, which is configured to:
analyze network traffic of a communication terminal of a user to determine a plurality of characteristics of the network traffic, wherein the traffic is of a plurality of unrecognized applications,
estimate, from the plurality of characteristics, a plurality of classes of the unrecognized applications that are installed on the communication terminal, without identifying the unrecognized applications,
exclude from the plurality of classes one or more of the plurality of classes corresponding to at least one of applications that are most commonly used or applications that are least-used,
create a combination of applications that are indicative of demographic attributes using the plurality of classes after excluding the one or more of the plurality of classes from the plurality of classes to remove the at least one of applications that are most commonly used or applications that are least-used from the combination of applications,
determine a respective usage pattern of each of the plurality of classes over time after excluding the one or more of the plurality of classes from the plurality of classes, and
dynamically deduce and update, by tracking added or removed applications or classes over time and using machine learning models, a demographic profile of the user of the communication terminal from the combination of applications and the respective usage pattern of each of the plurality of classes after excluding the one or more of the plurality of classes from the plurality of classes.
2. The system according to claim 1 , wherein the processor is configured to estimate the plurality of classes and deduce the demographic profile irrespective of content of the network traffic.
3. The system according to claim 1 , wherein the processor is configured to analyze the network traffic passively, without communicating with the communication terminal.
4. The system according to claim 1 , wherein at least part of the network traffic is encrypted.
5. The system according to claim 1 , wherein the processor is configured to obtain metadata of a recognized application from the application store, to assess a similarity between the metadata and attributes of one of the plurality of unrecognized applications installed on the communication terminal, and to classify the one of the plurality of unrecognized applications based on the similarity.
6. The system according to claim 1 , wherein the processor is configured to exclude at least one class prior to deducing the demographic profile.
7. The system according to claim 1 , wherein the demographic profile comprises one or more demographic attributes, and wherein the processor is configured to assign a confidence score to at least one of the demographic attributes.
8. The system according to claim 1 , wherein the demographic profile comprises at least one demographic attribute selected from a group of attributes consisting of age, gender, ethnic origin, mother tongue, marital status, education, occupation, employment, income level, number of people in household and residence type.
9. A method, comprising:
receiving network traffic from a mobile network relating to a communication terminal of a user;
analyzing the network traffic of the communication terminal of the user to determine a plurality of characteristics of the network traffic, wherein the traffic is of a plurality of unrecognized applications;
estimating, from the plurality of characteristics, a plurality of classes of the unrecognized applications that are installed on the communication terminal, without identifying the unrecognized applications;
excluding from the plurality of classes one or more of the plurality of classes corresponding to at least one of applications that are most commonly used or applications that are least-used;
creating a combination of applications that are indicative of demographic attributes using the plurality of classes after excluding the one or more of the plurality of classes from the plurality of classes to remove the at least one of applications that are most commonly used or applications that are least-used from the combination of applications;
determining a respective usage pattern of each of the plurality of classes over time after excluding the one or more of the plurality of classes from the plurality of classes; and
dynamically deducing and updating, by tracking added or removed applications or classes over time and using machine learning models, a demographic profile of the user of the communication terminal from the combination of applications and the respective usage pattern of each of the plurality of classes after excluding the one or more of the plurality of classes from the plurality of classes.
10. The method according to claim 9 , wherein analyzing the network traffic, estimating the plurality of classes, and deducing the demographic profile are performed irrespective of content of the network traffic.
11. The method according to claim 9 , wherein analyzing the network traffic is performed passively, without communicating with the communication terminal.
12. The method according to claim 9 , wherein at least part of the network traffic is encrypted.
13. The method according to claim 9 , wherein analyzing the network traffic comprises obtaining metadata of a recognized application from the application store, assessing a similarity between the metadata and attributes of one of the plurality of unrecognized applications installed on the communication terminal, and classifying the one of the plurality of unrecognized applications based on the similarity.
14. The method according to claim 9 , wherein analyzing the network traffic comprises excluding at least one class prior to deducing the demographic profile.
15. The method according to claim 9 , wherein the demographic profile comprises one or more demographic attributes, and wherein deducing the demographic profile comprises assigning a confidence score to at least one of the demographic attributes.
16. The method according to claim 9 , wherein the demographic profile comprises at least one demographic attribute selected from a group of attributes consisting of age, gender, ethnic origin, mother tongue, marital status, education, occupation, employment, income level, number of people in household and residence type.