Xad, Inc.

Yilin Chen

Can Liang

Pravesh Katyal

Amit Goswami

Jingqi Huang

Shun Jiang

Guoxin Li

Huitao Luo

Crystal Shi

Fan Yang

Weiqing Yu

Described herein are system and method for pacing information delivery to mobile devices. The method comprises, for each respective request of a first plurality of requests received during a time unit that qualifies for information delivery, predicting a respective conversion probability corresponding to a predicted probability of a mobile device associated with the respective request having at least one location event at any of one or more POIs during a time frame corresponding to the time unit. The method further comprises placing a bid for fulfilling the respective request based on the respective conversion probability and a bidding model, determining a set of predicted numbers of conversions corresponding, respectively, to a set of ranges of predicted conversion probabilities for a first number of fulfilled requests corresponding to the time unit, and adjusting the bidding model based at least on the predicted number of conversions.

Systems and methods for pacing information delivery to mobile devices

A system for predicting a conversion rate relating to targeted location events for a test campaign includes one or more campaign databases configured to store campaign parameters of a set of historical campaigns. The system further includes a feature engineering module configured to construct a training feature space corresponding to the set of historical campaigns, and to determine a set of labels including a respective conversion rate for each respective historical campaign of the plurality of historical campaigns. The system further includes a model training module configured to machine train a conversion rate prediction model using the training feature space and the set of labels. The feature engineering module is further configured to construct a set of test features from campaign parameters of the test campaign, and the system further includes a prediction module configured to apply the conversion rate prediction model to the set of test features to obtain a predicted conversion rate for the test campaign.

Systems and methods for predicting targeted location events

A system includes a machine learning module configured to train a location prediction model for an information campaign, a front-end server configured to receive and process information requests, and a prediction unit. During the information campaign, the prediction unit is configured to use the location prediction model to predict a conversion probability for any particular mobile device associated with a qualified information request received during any respective time unit. The conversion probability corresponds to a predicted probability of the particular mobile device having at least one location event at any of one or more POIs during a particular time frame. The front-end server is further configured to determine a respective target number of conversions to be achieved by the information campaign during the respective time unit, and to determine a response to the particular information request based at least in part on the conversion probability and on the respective target number of conversions.

Systems and methods for pacing information campaigns based on predicted and observed location events

A system includes a machine learning module configured to train a location prediction model using features constructed from mobile device data with time stamps in a training time period, and labels extracted from mobile device data with time stamps in a training time frame. The system further includes a prediction module configured apply the prediction model to a feature set constructed using mobile device data associated with a mobile device with time stamps in a prediction time period to obtain a prediction result corresponding to the mobile device. The system further includes a calibration module configured to obtain a calibration model corresponding to an information campaign, and a calibrated prediction module configured to apply the calibration model to the prediction result to obtain a calibrated probability for the mobile device to have at least one location event at any of one or more locations associated with the information campaign during a prediction time frame.

Systems and methods for calibrated location prediction

A system coupled to a packet-based network is configured to predict the locations of one or more mobile devices communicating with the packet-based network. The system comprises a request processor configured to detect location events associated with mobile devices communicating with the packet-based network, each location event corresponding to a time stamp and identifying a geo-place in a geo database. The system further comprises an off-line prediction subsystem configured to train a plurality of off-line prediction models and an on-line prediction model using various historical location events. The off-line prediction subsystem is further configured to generate off-line prediction results corresponding to the off-line prediction models. The system further comprises an on-line prediction subsystem configured to construct a feature vector using the off-line prediction results and recently detected location events in response to a request received in real-time, and to generate on-line prediction results by applying the on-line prediction model to the feature vector.

System and method for real-time prediction of mobile device locations

The present disclosure provides methods and systems that utilize mobile device location events and machine learning and generate predicative classification/regression model for lookalike prediction. Location related features, together with other user level information, are extracted, transformed and used as model feature input, and a client specified list of mobile devices or their associated users are used as prediction target. This system makes efficient use of different types of location events and thus offers improved scale and performance. It also enjoys many benefits offered by a machine learning platform, such as automatic adaptation to different lists of seed lists, addition of new features and changes in data statistical properties.

Systems and methods for predicting lookalike mobile devices

Systems and methods for using geo-blocks and geo-fences to discover lookalike mobile devices

Systems and methods for real-time prediction of mobile device locations

A system coupled to a packet-based network is configured to predict the locations of mobile devices that have communicated with the packet-based network. The system includes a request processor configured to detect location events associated with mobile devices communicating with the packet-based network, each location event corresponding to a time stamp and identifying a geo-place in a geo database. The geo-places include geo-blocks and geo-fences. The system further comprises a location prediction subsystem configured to construct first feature space using first location events and second feature space using second location events, and to extract a set of labels from third location events. The location prediction subsystem is further configured to train a prediction model using the first feature space and the set of labels, and to apply the prediction model to the second feature space to obtain prediction results.

Growing community of inventors

Sunnyvale, CA, United States of America

Yilin Chen