Identification of behavioral pattern of simulated transaction data

A computer program product for identifying a behavioral pattern from simulated transaction data generated on a cognitive system, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to:

• receive real transaction data of a plurality of real customers;
• receive statistical data representing the real transaction data;
• generate, by a cognitive system, simulated transaction data in imitation of the statistical data using a reinforcement learning model including an intelligent agent, a policy engine, and an environment;
• identify a behavioral pattern from the simulated transaction data;
• compare the behavioral pattern to the statistical data to determine whether the behavioral pattern is present in the statistical data;
• determine that the behavioral pattern is present in the statistical data;
• and
• train a fraudulent behavior detection model implemented on the cognitive system for determining whether a behavioral pattern indicates a fraudulent behavior using the simulated transaction data,
• wherein generating simulated transaction data further comprises:
• providing the statistical data as a goal;
• and
• performing a plurality of iterations to simulate the statistical data, wherein the plurality of iterations is performed until a degree of similarity of simulated transaction data relative to the statistical data is higher than a first predefined threshold such that the intelligent agent behaves similarly to the plurality of real customers;
• in each iteration:
• conducting, by the intelligent agent, an action including a plurality of simulated transactions;
• comparing, by the environment, the action with the goal;
• providing, by the environment, a feedback associated with the action based on a degree of similarity relative to the goal;
• and
• adjusting, by the policy engine, a policy based on the feedback, wherein the adjustment is configured to gain a superior feedback for a next action.

A system for identifying a behavioral pattern from simulated transaction data generated on a cognitive system, the system comprising:

• a processor configured to:
• receive real transaction data of a plurality of real customers;
• receive statistical data representing the real transaction data;
• generate, by a cognitive system, simulated transaction data in imitation of the statistical data using a reinforcement learning model including an intelligent agent, a policy engine, and an environment;
• identify a behavioral pattern from the simulated transaction data;
• compare the behavioral pattern with the statistical data to determine whether the behavioral pattern is present in the statistical data;
• determine that the behavioral pattern is present in the statistical data;
• and
• train a fraudulent behavior detection model implemented on the cognitive system for determining whether a behavioral pattern indicates a fraudulent behavior using the simulated transaction data,
• wherein generating simulated transaction data further comprises:
• providing the statistical data as a goal;
• and
• performing a plurality of iterations to simulate the statistical data, wherein the plurality of iterations is performed until a degree of similarity of simulated transaction data relative to the statistical data is higher than a first predefined threshold such that the intelligent agent behaves similarly to the plurality of real customers;
• in each iteration:
• conducting, by the intelligent agent, an action including a plurality of simulated transactions;
• comparing, by the environment, the action with the goal;
• providing, by the environment, a feedback associated with the action based on a degree of similarity relative to the goal;
• and
• adjusting, by the policy engine, a policy based on the feedback, wherein the adjustment is configured to gain a superior feedback for a next action.