US 12,393,854 B2
System and method for implementing an application prediction engine
Amar Srivastava, Morris Plains, NJ (US); Alexander N. Torres, Jersey City, NJ (US); Srinivasan Gopalakrishnan, Belle Mead, NJ (US); and Rohit Goldsmith, New York, NY (US)
Assigned to JPMORGAN CHASE BANK, N.A., New York, NY (US)
Filed by JPMorgan Chase Bank, N.A., New York, NY (US)
Filed on Apr. 12, 2021, as Appl. No. 17/227,919.
Claims priority of provisional application 63/008,027, filed on Apr. 10, 2020.
Prior Publication US 2021/0319349 A1, Oct. 14, 2021
Int. Cl. H04L 67/62 (2022.01); G06F 16/2455 (2019.01); G06F 16/903 (2019.01); G06F 16/957 (2019.01); G06N 7/01 (2023.01); G06N 20/00 (2019.01); H04L 67/306 (2022.01); H04L 67/5682 (2022.01)
CPC G06N 7/01 (2023.01) [G06F 16/90335 (2019.01); G06N 20/00 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A method that implements a prediction engine to provide predictive caching based on user behavior, the method comprising the steps of:
capturing, via a data collection processor, usage data and session data from a user via a client interface executing an application;
storing, via a prediction data storage, the usage data and session data;
based at least in part on the captured usage data and session data, generating, via a prediction engine, a transition probability matrix applying a forecasting request load model to data sequences associated with the user that identifies prediction data, wherein the usage data includes a click path navigation indicating a particular sequence of clicks providing a sequence matrix and timing of the clicks;
identifying an outlier based on the sequence matrix, and removing the outlier from the transition probability matrix to provide a modified transition probability matrix;
creating a schedule, based on the modified transition probability matrix, that identifies a timeframe at which the predicted data will be required by the user;
applying, via a task scheduler, the schedule to execute one or more tasks to cache data from an application database based on the schedule; and
retrieving, via a cache processor, data based on the schedule, wherein the cache processor communicates with the user via a service layer associated with the application that executes a query service and a cache service based on a user request.