Venue detection
Examples of the present disclosure describe systems and methods for venue detection. In aspects, a mobile device comprising a set of sensors may collect and store sensor data from in response to a detected movement event or user interaction data. The sensor data may be used to generate a set of candidate venues corresponding to the location of the mobile device. The candidate venues may be provided to a venue detection system. The venue detection system may process the candidate venues to generate a set of features. The set of features may be applied to, and/or used to generate, one or more probabilistic models. The probabilistic models may generate confidence metrics for each of the candidate venues. In some aspects, the top ‘N’ venues may be selected from the set. The top ‘N’ venues may then be presented to the user and/or used to effect one or more actions.
1. A method for venue detection, the method comprising:
receiving sensor data from a mobile device, the sensor data is associated with a location of the mobile device, wherein the sensor data is collected in response to detection of an event;
generating a venue list using the sensor data, wherein the venue list comprises at least a first venue and a second venue, wherein the first venue and the second venue are associated with the location of the mobile device;
featurizing venue data associated with the first venue and the second venue to generate a feature set;
generating confidence metrics for the first venue and the second venue by applying the feature set to a probabilistic model;
calibrating the venue list based in part on the metrics;
presenting the calibrated venue list and at least a portion of the confidence metrics via a user interface, wherein at least one of the first venue and the second venue corresponds to the location of the mobile device;
providing output data generated by the probabilistic model to an objective function, wherein the probabilistic model is an ensemble of decision trees generated using one or more gradient boosting techniques, and wherein the objective function defines a number and type of additional probabilistic models to evaluate against the probabilistic model;
using the objective function to compare the output data to training data comprising previously received input data and corresponding result data; and
causing at least a subset of the output data to be rendered for display on the mobile device.
2. The method of claim 1 , wherein the one or more gradient boosting techniques evaluate at least one of: venue age, venue popularity, proximity to other venues, historical accuracy of venue candidates, previous venue visit data, and user feedback.
3. The method of claim 1 , wherein the sensor data is indicative of a visit by the mobile device to at least one of the first venue and the second venue.
4. The method of claim 1 , further comprising:
based on comparing the output data generated by the probabilistic model to the output data in the training data, scoring the effectiveness of the probabilistic model.
5. The method of claim 3 , wherein the confidence metrics indicate a probability a respective venue corresponds to the location of the mobile device.
6. The method of claim 1 , wherein the generating the venue list comprises determining a first set of venues in a first proximity to the location of the mobile device, and wherein calibrating the venue list comprises determining a second set of venues in a second proximity to the location of the mobile device, wherein the second set of venues comprises fewer venues than the first set of venues.
7. A non-transitory computer storage media encoding computer executable instructions which, when executed by at least one processor, performs a method for venue detection, the method comprising:
receiving sensor data from a mobile device, the sensor data is associated with a location of the mobile device, wherein the sensor data is collected in response to detection of an event;
generating a venue list using the sensor data, wherein the venue list comprises at least a first venue and a second venue, wherein the first venue and the second venue are associated with the location of the mobile device;
featurizing venue data associated with the first venue and the second venue to generate a feature set;
generating confidence metrics for the first venue and the second venue by applying the feature set to a probabilistic model;
calibrating the venue list based in part on the metrics;
providing output data generated by the probabilistic model to an objective function, wherein the probabilistic model is an ensemble of decision trees generated using one or more gradient boosting techniques, and wherein the objective function defines a number and type of additional probabilistic models to evaluate against the probabilistic model;
using the objective function to compare the output data to training data comprising previously received input data and corresponding result data; and
causing at least a subset of the output data to be rendered for display on the mobile device.
8. The non-transitory computer storage media of claim 7 , wherein the sensor data is indicative of a visit by the mobile device to the one or more venues.
9. The non-transitory computer storage media of claim 7 , wherein the one or more gradient boosting techniques evaluate at least one of: venue age, venue popularity, proximity to other venues, historical accuracy of venue candidates, previous venue visit data, and user feedback.
10. The non-transitory computer storage media of claim 7 , wherein the generating the venue list comprises determining a first set of venues in a first proximity to the location of the mobile device, and wherein calibrating the venue list comprises determining a second set of venues in a second proximity to the location of the mobile device, wherein the second set of venues comprises fewer venues than the first set of venues.
11. A system comprising:
one or more processors; and
memory coupled to at least one of the one or more processors, the memory comprising computer executable instructions that, when executed by the at least one processor, performs a method for venue detection, the method comprising:
receiving sensor data from a mobile device, the sensor data is associated with a location of the mobile device, wherein the sensor data is collected in response to detection of an event;
generating a venue list using the sensor data, wherein the venue list comprises one or more candidate venues corresponding to the location of the mobile device;
featurizing venue data associated with the one or more candidate venues to generate a feature set;
generating metrics for the one or more candidate venues by applying the feature set to a probabilistic model;
calibrating the venue list based in part on the metrics;
presenting the calibrated venue list using a user interface;
receiving, via the user interface, user input associated with the calibrated venue list; and
using the user input to modify the calibrated venue list,
providing output data generated by the probabilistic model to an objective function, wherein the objective function defines a number and type of additional probabilistic models to evaluate against the probabilistic model;
using the objective function to compare the output data to training data comprising previously received input data and corresponding result data; and
causing at least a subset of the output data to be rendered for display on the mobile device.
12. The system of claim 11 , wherein the sensor data comprises at least one of:
geolocation data, positional data, Wi-Fi information, software information, hardware information, accelerometer data and time information.
13. The system of claim 11 , wherein the sensor data is indicative of a visit by the mobile device to one or more venues.
14. The system of claim 11 , wherein featurizing the venue data comprises:
identifying one or more features in the venue data; and
using the one or more features to generate the feature set.
15. The system of claim 14 , wherein identifying one or more features comprises evaluating at least one of: venue age, venue popularity, proximity to other venues, historical accuracy of venue candidates and previous venue visit data.
16. The system of claim 11 , wherein the probabilistic model is an ensemble of decision trees generated using one or more gradient boosting techniques.
17. The system of claim 16 , wherein the probabilistic model determines a confidence score for the one or more candidate venues, wherein the confidence score indicates a probability a corresponding venue corresponds to the location of the mobile device.
18. The system of claim 11 , wherein generating the venue list comprises at least one of: identifying venue information within the sensor data, and providing a set of geographical coordinates representative of the location to a venue determination utility.
19. The system of claim 11 , wherein calibrating the venue list comprises selecting a top ‘N’ venues, wherein the top ‘N’ venues represent one or more venues most likely to correspond to the location of a mobile device.
20. The system of claim 11 , wherein the sensor data is received in response to detecting at least one of a stop event by the mobile device and a visit event by the mobile device.