| CPC G06Q 30/0224 (2013.01) [G06N 5/04 (2013.01); G06N 20/00 (2019.01); G06Q 30/0235 (2013.01)] | 20 Claims |
|
1. A method, comprising:
generating, by one or more computing devices associated with an enterprise and by applying an offer decay generation model to an indication of a first customer, a first offer decay curve that is optimized for the first customer and that is for an incentive corresponding to a product provided by the enterprise,
the offer decay generation model being a machine learning model trained on historical data indicative of a plurality of interactions of a plurality of customers with the enterprise, a plurality of products provided by the enterprise, a plurality of offered incentives associated with the plurality of products, and a plurality of outcomes of the plurality of offered incentives to determine offer decay curves that are specific to individual customers and that are configured to maximize probabilities of the individual customers purchasing products; and
the first offer decay curve indicating an ordered set of decreasing values of the incentive, each value of the ordered set of decreasing values corresponding to a respective subsequent time interval during which the each value is valid;
updating the historical data to include an indication of the first offer decay curve and an indication of a response, of the first customer, to the incentive;
re-training the offer decay generation model based on the updated historical data; and
generating, by applying the re-trained offer decay generation model to another indication of the first customer or to an indication of a second customer, a second offer decay curve that is optimized for the first customer or for the second customer.
|