Recurrent neural network for controlling errors in data processing performance levels
AssigneeTruist Bank
InventorsMatthew Laine Donlan, Michael Joseph Senuta
Life of this patent
Estimating the term…
Loading patent…
AssigneeTruist Bank
InventorsMatthew Laine Donlan, Michael Joseph Senuta
Estimating the term…
A method and a computing system evaluate and control a performance level of an automated data processing operation being performed. The system: receives real-time operational performance data for the operation being performed; evaluates the data and generates a current performance level for the operation; trains a recurrent neural network to build a machine learning model to process the received operational performance data and generate a predicted performance level for the operation corresponding to a predetermined future time; generates an error detection indicator representing a difference between the predicted performance level and a specified performance level; displays information representing the specified, the current and the predicted performance levels and the error detection indicator on a display; and controls the operation to correct the current performance level to the specified performance level when the error detection indicator exceeds a specified performance difference between the current performance level and the predetermined performance level.
A computing system configured to evaluate and control a data processing performance level of an automated data processing operation being performed, the computing system comprising:
A method using a recurrent neural network to evaluate and control a data processing performance level of an automated data processing operation, the method comprising steps of:
Dependencies are inferred from the claim text.