Systems and methods for on-the-fly alert triggering customization

Abstract

Systems and methods for feature-based alert triggering are disclosed herein. The system can include memory including a model database containing a machine-learning algorithm. The system can include a user device that can receive inputs from a user; and at least one server. The at least one server can: receive electrical signals from the user device, the electrical signals corresponding to a plurality of user inputs provided to the user device; automatically generate input-based features from the received electrical signals; input the input-based features into the machine-learning algorithm; automatically and directly generate a risk prediction with the machine-learning algorithm from the input-based features; and generate and display an alert when the risk prediction exceeds a threshold value.

What is claimed is: 
     
         1 . A system for on-the-fly alert triggering customization, the system comprising:
 memory comprising:
 a machine-learning classifier configured to generate a risk prediction based on inputted features; 
 a user profile database identifying a user and containing metadata associated with the user; and 
 a customization database identifying one or several user attributes and a customization associated with each of those one or several user attributes, wherein the customization identifies a sub-set of potential features for use in generating a risk prediction; 
   a first user device configured to receive inputs from a user;   a second user device configured to display information to a user; and   at least one server configured to:
 receive electrical signals corresponding to user inputs to the first user device; 
 retrieve metadata associated with the user of the first user device; 
 identify a customization for the user of the first user device based on the retrieved metadata; 
 select a sub-set of input-based features from the received electrical signals according to the identified customization; 
 input the sub-set of the features into the machine-learning classifier; and 
 generate a customized risk prediction with the machine-learning classifier. 
   
     
     
         2 . The system of  claim 1 , wherein the at least one server is further configured to modify the machine-learning classifier according to the identified customization. 
     
     
         3 . The system of  claim 2 , wherein the machine-learning classifier comprises a plurality of classifiers, wherein each of the plurality of classifiers is associated with a unique set of features. 
     
     
         4 . The system of  claim 3 , wherein modifying the machine-learning classifier comprises selecting one of the plurality of classifiers corresponding to the sub-set of features selected according to the customization. 
     
     
         5 . The system of  claim 1 , wherein the at least one server is further configured to control the second user device to display an alert when the risk prediction exceeds a threshold value. 
     
     
         6 . The system of  claim 5 , wherein the alert comprises a graphical depiction of the risk prediction. 
     
     
         7 . The system of  claim 1 , wherein the metadata is unique to the user. 
     
     
         8 . The system of  claim 1 , wherein the customization is determined according to a portion of the metadata that is non-unique to the user and is unique to a set of users sharing at least one common attribute. 
     
     
         9 . The system of  claim 1 , wherein inputting the sub-set of the features into the machine-learning classifier comprises: generating a feature vector for each of the features in the sub-set of features; and inputting the feature vectors into the classifier. 
     
     
         10 . The system of  claim 1 , wherein the at least one server is further configured to generate a set of features, and wherein the sub-set of features is selected from the generated set of features. 
     
     
         11 . A method for on-the-fly alert triggering customization, the method comprising:
 receiving electrical signals corresponding to user inputs to a first user device;   retrieving metadata associated with a user of the first user device;   identifying a customization for the user of the first user device based on the retrieved metadata;   selecting a sub-set of input-based features from the received electrical signals according to the identified customization;   inputting the sub-set of the features into a machine-learning classifier; and   generating a customized risk prediction with the machine-learning classifier.   
     
     
         12 . The method of  claim 11 , further comprising modifying the machine-learning classifier according to the identified customization. 
     
     
         13 . The method of  claim 12 , wherein machine-learning classifier comprises a plurality of classifiers, wherein each of the plurality of classifiers is associated with a unique set of features. 
     
     
         14 . The method of  claim 13 , wherein modifying the machine-learning classifier comprises selecting one of the plurality of classifiers corresponding to the sub-set of features selected according to the customization. 
     
     
         15 . The method of  claim 11 , further comprising controlling a second user device to display an alert when the risk prediction exceeds a threshold value. 
     
     
         16 . The method of  claim 15 , wherein the alert comprises a graphical depiction of the risk prediction. 
     
     
         17 . The method of  claim 11 , wherein the metadata is unique to the user. 
     
     
         18 . The method of  claim 11 , wherein customization is determined according to a portion of the metadata that is non-unique to the user and that is unique to a set of users sharing at least one common attribute. 
     
     
         19 . The method of  claim 11 , wherein inputting the sub-set of the features into the machine-learning classifier comprises: generating a feature vector for each of the features in the sub-set of features; and inputting the feature vectors into the classifier. 
     
     
         20 . The method of  claim 11 , further comprising generating a set of features, wherein the sub-set of features is selected from the generated set of features.