Fraud prevention in programmatic advertising

Abstract

Embodiments of the present invention provide for machine learning-based systems and methods for preventing fraud in programmatic advertising. The systems and methods provide for applying a plurality of machine learning models to data associated with a bid request, determining if the bid request is associated with fraudulent activity as a result of the machine learning models, and selectively preventing the bid request from being provided to potential buyers based on the determination.

1 - 20  (canceled) 
     
     
         21 . A computer-implemented method for preventing fraud in programmatic advertising, the method comprising:
 applying first and second sets of machine learning models to data associated with a bid request, wherein the second set of machine learning models is different from the first set of machine learning models, wherein the first and second sets of machine learning models are: (i) trained on respective training data sets and (ii) re-trained on updated information, wherein at least one of the first and second sets of machine learning models is configured to predict at least one fraudulent activity; and   determining if the bid request is associated with the at least one fraudulent activity based on a result from applying at least one of the first set of machine learning models and the second set of machine learning models to the data associated with the bid request.   
     
     
         22 . The method of  claim 21 , further comprising:
 selectively preventing the bid request from being provided to at least one buyer based on the determination.   
     
     
         23 . The method of  claim 21 , wherein the data associated with the bid request is classified by data types, wherein the data types include string, integer, and float. 
     
     
         24 . The method of  claim 21 , wherein the first set of machine learning models includes: (i) a machine learning model testing for unusual volume spikes and (ii) a machine learning model testing for a randomness factor. 
     
     
         25 . The method of  claim 21 , wherein the second set of machine learning models includes: (i) a machine learning model testing for synthetically-generated users and (ii) a machine learning model testing for data reset fraud. 
     
     
         26 . The method of  claim 21 , wherein the bid request is associated with at least one advertising inventory on a one of a website, a mobile application, and an Internet-connected device. 
     
     
         27 . The method of  claim 26 , wherein the at least one advertising inventory is a display ad on the one of the website, the mobile application, and the Internet-connected device. 
     
     
         28 . The method of  claim 21 , further comprising:
 applying at least one validation test to the data associated with the bid request, wherein the at least one validation test includes at least one of an obsolete data test, an inaccurate data test, and an incompatible data test;   determining if the bid request is associated with the at least one fraudulent activity based on a result from applying the at least one validation test; and   selectively preventing the bid request from being provided to the at least one buyer based on the determination.   
     
     
         29 . The method of  claim 21 , further comprising:
 comparing the data associated with the bid request to a list of known fraudulent domains and/or applications;   determining if the bid request is associated with the at least one fraudulent activity based on the comparison; and   selectively preventing the bid request from being provided to the at least one buyer based on the determination.   
     
     
         30 . A computer-implemented system for preventing fraud in programmatic advertising, the system comprising:
 at least one server comprising a memory, a processor, and a representational state transfer application programming interface (RESTful API), wherein the at least one server is configured to:
 apply first and second sets of machine learning models to data associated with a bid request, wherein the second set of machine learning models is different from the first set of machine learning models, wherein the first and second sets of machine learning models are: (i) trained on respective training data sets and (ii) re-trained on updated information, wherein at least one of the first and second sets of machine learning models is configured to predict at least one fraudulent activity; and 
 determine if the bid request is associated with the at least one fraudulent activity based on a result from applying at least one of the first set of machine learning models and the second set of machine learning models to the data associated with the bid request. 
   
     
     
         31 . The system of  claim 30 , wherein the bid request is associated with at least one advertising inventory on a website, a mobile application, and an Internet-connected device. 
     
     
         32 . The system of  claim 30 , wherein the at least one advertising inventory is a display ad on the one of the website, the mobile application, and the Internet-connected device. 
     
     
         33 . The system of  claim 30 , wherein the bid request is transmitted from another server. 
     
     
         34 . The system of  claim 30 , wherein the RESTful API is configured to run on a machine port of a computer. 
     
     
         35 . The system of  claim 34 , wherein the RESTful API is configured to receive the bid request on the machine port. 
     
     
         36 . The system of  claim 35 , wherein the RESTful API is configured to receive the bid request from at least one other application on the computer.