Passive visit detection

Abstract

Examples of the present disclosure describe systems and methods for passive visit detection. In aspects, a mobile device comprising a set of sensors may collect and store sensor data from the set of sensors in response to detecting a movement event or user interaction data. The collected sensor data may be processed and provided as input to one or more predictive or statistical models. The model(s) may evaluate the sensor data to detect mobile device location, movement events and visit events. The model(s) may also be used to determine correlations between features of the sensor data and movement- or location-based events, optimize the types of data collected by the set of sensors, extend localized predictions to large-scale ecosystems, and generate battery-efficient state predictions. In aspects, the model(s) may be trained using labeled and/or unlabeled data sets of sensor data.

What is claimed is: 
     
         1 . 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 passive visit detection, the method comprising:
 receiving sensor data for a mobile device, wherein the sensor data comprises at least movement data for the mobile device and network data for a venue; 
 processing the sensor data to identify a feature set; and 
 evaluating the feature set to determine whether a visit event has occurred. 
   
     
     
         2 . The system of  claim 1 , the method further comprising:
 detecting, by the mobile device, one or more events corresponding to the sensor data, wherein the one or more events correspond to at least one of movement events, purchase events, information delivery events and venue check-in events; and   correlating the one or more events to the one or more features.   
     
     
         3 . The system of  claim 1 , wherein the movement data comprises geographic location data, and the network data comprises one or more Wi-Fi signal strengths for the venue. 
     
     
         4 . The system of  claim 1 , wherein processing the sensor data comprises:
 parsing the sensor data to identify the one or more Wi-Fi signal strengths;   using a statistical model to evaluate the identified one or more Wi-Fi signal strengths; and   generating, by the statistical model, at least a portion of the feature set.   
     
     
         5 . The system of  claim 4 , wherein evaluating the one or more Wi-Fi signal strengths comprises using the Wi-Fi signal strengths to determine a distance between the mobile device and the venue. 
     
     
         6 . The system of  claim 4 , wherein evaluating the one or more Wi-Fi signal strengths comprises:
 observing a first set of network signals at a first time;   observing a second set of network signals at a second time; and   comparing the first set of network signals to the second set of network signals to determine an estimated distance traveled by the mobile device between the first time and the second time.   
     
     
         7 . The system of  claim 1 , wherein evaluating the feature set comprises providing the feature set to a predictive model operable to detect a visit state of the mobile device for one or more venues. 
     
     
         8 . The system of  claim 7 , wherein the predictive model is an HMM operable to determine features, in the feature sets, relevant to detecting a visit event. 
     
     
         9 . The system of  claim 8 , wherein the features determined using the HMM are used as an initialization point for an EM algorithm operable to evaluate a set of unlabeled data. 
     
     
         10 . The system of  claim 8 , wherein the predictive model is implemented on the mobile device, and the predictive model is operable to detect visit states in real time. 
     
     
         11 . The system of  claim 8 , wherein the predictive model is implemented on a remote device operable to train the predictive model using labeled data and perform big data analysis. 
     
     
         12 . The system of  claim 1 , wherein the method further comprises generating a set of observations for the sensor data, wherein the set of observations are indicative of a correlation between one or more features of the sensor data and a visit state. 
     
     
         13 . The system of  claim 1 , wherein the system further comprises a user interface operable to receive user input comprising at least one of a data label and training data. 
     
     
         14 . A method for passive visit detection, the method comprising:
 receiving, at a mobile device, sensor data, wherein the sensor data comprises at least one of: location data for the mobile device, movement data for the mobile device, and network data for a venue;   processing the sensor data in real time to generate a feature set for the sensor data; and   using an HMM implementation to evaluate the feature set to determine a visit state of the mobile device.   
     
     
         15 . The method of  claim 14 , wherein the sensor data comprises telemetry data for a first time period and at least a second time period, and wherein evaluating the feature set comprises comparing sensor data for the first time period to sensor data for the second time period. 
     
     
         16 . The method of  claim 15 , wherein evaluating the feature set further comprises correlating features in the feature set to one or more events, wherein the one or more events correspond to at least one of movement events, purchase events, information delivery events and venue check-in events. 
     
     
         17 . A method for passive visit detection, the method comprising:
 receiving, at a computing device, sensor data, wherein the sensor data comprises at least one of: location data for a mobile device, movement data for the mobile device, and network data for a venue;   processing the sensor data to generate a feature set for the sensor data; and   using the feature set to train a predictive model to detect one or more visit states of the mobile device.   
     
     
         18 . The method of  claim 17 , wherein the sensor data comprises a set of labeled data, and wherein the predictive model uses the set of labeled data to determine correlations between the set of labeled data and the one or more visit states. 
     
     
         19 . The method of  claim 18 , wherein the determined correlations are stored in one or more data stores for use with subsequent training of the predictive model. 
     
     
         20 . The method of  claim 17 , wherein an instance of the trained predictive model is transmitted to the mobile device.