System and method for segmenting mobile entities based on detected mobile events

1. A method for segmenting mobile entities, comprising:

at one or more computer systems coupled to a packet-based network and including or having access to electronic storage media storing therein events associated with mobile entities communicating with the packet-based network using signals, each of the events having one or more features;

determining a first set of mobile entities and a second set of mobile entities based at least on predefined constraints including first constraints and second constraints, each mobile entity in the first set of mobile entities satisfying the first constraints and the second constraints, each mobile entity in the second set of mobile entities satisfying the first constraints but not the second constraints, wherein the first constraints include a condition that, for each location of a set of predefined locations, a frequency of a mobile entity being detected at the each location is less than a preset maximum frequency, and wherein the second constraints include a condition that a frequency of a mobile entity having been detected at any of the set of predefined locations is greater than or equal to a preset overall frequency;

identifying a feature set related to a mobile segment, the feature set including a plurality of features;

determining a plurality of feature gains corresponding, respectively, to the plurality of features related to the mobile segment, wherein a respective feature gain corresponding to a respective feature is related to a respective first probability of finding the respective feature in any of a set of events associated with mobile entities in the first set of mobile entities and a second probability of finding the respective feature in any of a set of events associated with mobile entities in the second set of mobile entities; and

determine a set of mobile entities to add to the mobile segment based at least on the plurality of feature gains, and for each particular feature of the plurality of features, frequency of events associated with each of the set of mobile entities and having the particular feature.

2. The method of claim 1 , further comprising:

assigning weights to the plurality of features, wherein the plurality of features include one or more location features corresponding to one or more geo-fences, and one or more non-location features, and wherein the weight assigned to any of the one or more location features is greater than a weight assigned to any of the one or more non-location features;

wherein the set of mobile entities are further determined based on the weights assigned to the plurality of features.

3. The method of claim 2 , wherein:

the location features include one or more first location features each corresponding to at least one geo-fence of a first type, and one or more second location features each corresponding to at least one geo-fence of a second type, and wherein each of the one or more first location features is assigned a weight that is greater than a weight assigned to any of the one or more second location features.

4. The method of claim 3 , wherein the location features include brand names associated with points of interest (POI's).

5. The method of claim 3 , wherein each of the at least one geo-fence of the first type has a polygonal shape outlining a building or a premise of a point of interest (POI), and wherein each of the at least one geo-fence of the second type has a circular shape with a predefined radius.

6. The method of claim 2 , wherein the one or more non-location features include at least one feature related to an application operable on mobile devices.

7. The method of claim 1 , wherein each respective mobile entity of the set of mobile entities is added to the segment based on a weighted sum of a plurality of terms corresponding, respectively, to at least some of the plurality of features, each specific term of the plurality of terms being related to a feature gain of a corresponding feature weighted by a weight assigned to the corresponding feature and frequency of events associated with the respective mobile entity and having the corresponding feature.

8. The method of claim 1 , wherein the plurality of events further include feedback events, and the plurality of factors further include one or more measures corresponding to one or more types of feedback events associated with the respective mobile entity, the one or more types of feedback events include one or more of a type of feedback events associated with impressions of certain information on mobile devices, a type of feedback events associated with clicks on certain links displayed on mobile devices, a type of feedback events associated with calls to certain phone numbers made with mobile devices, and a type of feedback events associated with secondary actions carried out using mobile devices.

9. The method of claim 1 , wherein the one or more predefined locations include multiple locations associated with multiple brands related to the mobile segment.

10. The method of claim 9 , wherein the preset maximum frequency is much greater than the preset overall frequency.

11. A method for discovering mobile entities to form a mobile segment that includes a plurality of brands, comprising:

at one or more computer systems coupled to a packet-based network and including or having access to electronic storage media storing therein events associated with mobile entities communicating with the packet-based network using signals, each of the events having one or more features;

determining a first set of mobile entities and a second set of mobile entities based at least on predefined constraints including first constraints and second constraints, each mobile entity in the first set of mobile entities satisfying the first constraints and the second constraints, each mobile entity in the second set of mobile entities satisfying the first constraints but not the second constraints, wherein the first constraints includes a condition that, for each brand of the plurality of brands, a frequency of events indicating a mobile entity being at any of one or more locations associated with the each brand is less than a preset maximum frequency, and wherein the second constraints includes a condition that a frequency of events indicating a mobile entity at any of a predetermined set of locations is greater than or equal to a preset overall frequency, each of the plurality of brands having at least one associated location in the predetermined set of locations;

identifying a feature set related to the mobile segment, the feature set including a plurality of features;

determining a plurality of feature gains corresponding, respectively, to the plurality of features related to the mobile segment, wherein a respective feature gain corresponding to a respective feature is related to a respective first probability of finding the respective feature in any of a set of events associated with mobile entities in the first set of mobile entities and a second probability of finding the respective feature in any of a set of events associated with mobile entities in the second set of mobile entities; and

determine a set of mobile entities to add to the mobile segment based at least on the plurality of feature gains, and for each particular feature of the plurality of features, frequency of events associated with each of the set of mobile entities and having the particular feature.

12. The method of claim 11 , further comprising:

assigning weights to the plurality of features, wherein the plurality of features include one or more location features corresponding to one or more geo-fences, and one or more non-location features, and wherein the weight assigned to any of the one or more location features is greater than a weight assigned to any of the one or more non-location features;

wherein the set of mobile entities are further determined based on the weights assigned to the plurality of features.

13. The method of claim 12 , wherein:

the location features include one or more first location features each corresponding to at least one geo-fence of a first type, and one or more second location features each corresponding to at least one geo-fence of a second type, and wherein each of the one or more first location features is assigned a weight that is greater than a weight assigned to any of the one or more second location features.

14. The method of claim 13 , wherein the location features include brand names associated with points of interest (POI's).

15. The method of claim 13 , wherein each of the at least one geo-fence of the first type has a polygonal shape outlining a building or a premise of a point of interest (POI), and wherein each of the at least one geo-fence of the second type has a circular shape with a predefined radius.

16. The method of claim 12 , wherein the one or more non-location features include at least one feature related to an application operable on mobile devices.

17. The method of claim 11 , wherein each respective mobile entity of the set of mobile entities is added to the segment based on a weighted sum of a plurality of terms corresponding, respectively, to at least some of the plurality of features, each specific term of the plurality of terms being related to a feature gain of a corresponding feature weighted by a weight assigned to the corresponding feature and frequency of events associated with the respective mobile entity and having the corresponding feature.

18. The method of claim 11 , wherein the plurality of events further include feedback events, and the plurality of factors further include one or more measures corresponding to one or more types of feedback events associated with the respective mobile entity, the one or more types of feedback events include one or more of a type of feedback events associated with impressions of certain information on mobile devices, a type of feedback events associated with clicks on certain links displayed on mobile devices, a type of feedback events associated with calls to certain phone numbers made with mobile devices, and a type of feedback events associated with secondary actions carried out using mobile devices.

19. A system for discovering mobile entities to form a segment that includes a plurality of brands, comprising:

electronic storage media storing therein events associated with mobile entities communicating with the packet-based network using signals, each of the events having one or more features;

one or more processors in one or more computer systems coupled to a packet-based network and including or having access to the electronic storage;

memory coupled to the one or more processors and storing therein program instructions, which, when executed by the one or more processors, cause the one or more processors to carry out a method, comprising:

determining a first set of mobile entities and a second set of mobile entities based at least on predefined constraints including first constraints and second constraints, each mobile entity in the first set of mobile entities satisfying the first constraints and the second constraints, each mobile entity in the second set of mobile entities satisfying the first constraints but not the second constraints, wherein the first constraints includes a condition that, for each brand of the plurality of brands, a frequency of events indicating a mobile entity being at any of one or more locations associated with the each brand is less than a preset maximum frequency, and wherein the second constraints includes a condition that a frequency of events indicating a mobile entity at any of a predetermined set of locations is greater than or equal to a preset overall frequency, each of the plurality of brands having at least one associated location in the predetermined set of locations;

identifying a feature set related to a mobile segment, the feature set including a plurality of features;

determining a plurality of feature gains corresponding, respectively, to the plurality of features related to the mobile segment, wherein a respective feature gain corresponding to a respective feature is related to a respective first probability of finding the respective feature in any of a set of events associated with mobile entities in the first set of mobile entities and a second probability of finding the respective feature in any of a set of events associated with mobile entities in the second set of mobile entities; and

determine a set of mobile entities to add to the mobile segment based at least on the plurality of feature gains, and for each particular feature of the plurality of features, frequency of events associated with each of the set of mobile entities and having the particular feature.

20. The system of claim 19 , wherein each respective mobile entity of the set of mobile entities is added to the segment based on a weighted sum of a plurality of terms corresponding, respectively, to at least some of the plurality of features, each specific term of the plurality of terms being related to a feature gain of a corresponding feature weighted by a weight assigned to the corresponding feature and frequency of events associated with the respective mobile entity and having the corresponding feature.