US 12,314,976 B2
System, method, and computer program product for generating a synthetic control group
Pulkit Aggarwal, Mountain View, CA (US); Paul Max Payton, San Carlos, CA (US); Lace Cheung, San Francisco, CA (US); and Suresh Krishna Vaidyanathan, Dublin, CA (US)
Assigned to Visa International Service Association, San Francisco, CA (US)
Filed by Visa International Service Association, San Francisco, CA (US)
Filed on Aug. 28, 2023, as Appl. No. 18/456,669.
Application 18/456,669 is a continuation of application No. 17/697,183, filed on Mar. 17, 2022, granted, now 11,776,005.
Application 17/697,183 is a continuation of application No. 16/801,653, filed on Feb. 26, 2020, granted, now 11,308,515, issued on Apr. 19, 2022.
Prior Publication US 2023/0401601 A1, Dec. 14, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06Q 30/0242 (2023.01); G06Q 30/0201 (2023.01); G06Q 30/0202 (2023.01)
CPC G06Q 30/0244 (2013.01) [G06Q 30/0201 (2013.01); G06Q 30/0202 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
after at least one holder of at least one transaction account of a plurality of transaction accounts is exposed to at least one communication via a computer-implemented advertisement program in a first time period, receiving, with at least one processor, transaction data associated with at least one transaction completed by a first set of transaction accounts of the plurality of transaction accounts with at least one target merchant in the first time period, the transaction data generated by a transaction service provider system from processing the at least one transaction;
generating, with the at least one processor, a synthetic control group comprising a subset of transaction accounts sampled from the first set of transaction accounts;
determining, with the at least one processor, for each transaction account of the synthetic control group, a propensity score representative of a likelihood of said transaction account being associated with a test group using a first machine learning model, wherein holders of transaction accounts in the test group were exposed to the at least one communication;
determining, with the at least one processor, for each transaction account of the synthetic control group, a predictive spending score for a second time period using a second machine learning model;
balancing, with the at least one processor, the synthetic control group, wherein balancing the synthetic control group comprises:
assigning an entropy balancing weight to each transaction account of the synthetic control group based at least partly on historic transaction data, the propensity score of each transaction account of the synthetic control group, and the predictive spending score of each transaction account of the synthetic control group;
reducing, with the at least one processor, computing resources required for the computer-implemented advertisement program in the second time period subsequent the first time period by altering at least one operational parameter of the computer-implemented advertisement program based on the transaction data and the synthetic control group, wherein the at least one operational parameter comprises a time of communications to be transmitted; and
executing, with the at least one processor, the computer-implemented advertisement program in the second time period based on the at least one operational parameter that was altered, wherein executing the computer-implemented advertisement program comprises:
allocating a reduced amount of processing capacity and memory storage space to execute the computer-implemented advertisement program based on the at least one operational parameter that was altered and further based on server uptime.