| CPC G06Q 30/0236 (2013.01) [G06Q 30/0212 (2013.01); G06Q 30/0224 (2013.01); G06N 20/00 (2019.01)] | 9 Claims |
|
1. A system for training a machine learning model and automatically updating an account setting, the system comprising:
a computer with one or more processor and memory, wherein the computer executes computer-readable instructions, and wherein the computer is managed by a business entity; and
wherein, upon execution of the computer-readable instructions, the computer performs steps comprising:
requesting an initial response to one of a plurality of initial queries from a plurality of users, wherein each of the users has an account with the business entity, wherein the account of each of the users is accessible via a software application and/or a website managed by the computer of the business entity, wherein the account of each of the users is associated with a plurality of account settings that are selectable by each of the respective users, wherein each of the account settings is associated with a manner in which the computer interacts with a respective user device of each of the respective users via a visually perceptible aspect of a graphical user interface of the software application and/or the website managed by the computer of the business entity, wherein the initial response is requested when each of the respective users is enrolling or participating in a first stage of a multi-stage contest, wherein eligibility for a prize of the multi-stage contest requires each corresponding user providing the initial response to the one of the plurality of the initial queries, wherein the plurality of initial queries includes a first initial query and a second initial query different from the first initial query, wherein the first initial query includes an inquiry of a relationship between the multi-stage contest and a first financial product and/or service that is offered by the business entity managing the computer and the business entity is a financial institution, wherein the plurality of users includes a plurality of first users requested to provide the initial response to the first initial query and not the second initial query, a plurality of second users requested to provide the initial response to the second initial query and not the first initial query, and a plurality of third users that are not requested to provide the initial response to either of the first initial query or the second initial query, wherein the second query is related to a preference of each of the responding first users or responding second users regarding a selection of one of the account settings, wherein the initial response to the first initial query includes information related to a preference for the first financial product and/or service, wherein the initial response to the second initial query includes information about the one of the account settings;
storing the initial response of each of the responding users as initial response data, the initial response data forming a subset of a personal data set of each of the responding users;
generating an initial predictive model during initial training of a machine learning program utilizing at least one neural network, an initial training data set utilized during the initial training of the machine learning program comprising the personal data set of each of the plurality of users following the storing of the initial response of each of the responding users as initial response data;
deploying the trained initial predictive model;
predicting, by the initial predictive model, a predicted second response of each of the first users to the second initial query and predicting a predicted first response of each of the plurality of second users to the first initial query, the predicted first response indicating a likelihood that each of the plurality of second users will obtain the first financial product and/or service by correlating a respective subset of personal data of the plurality of second users to a respective subset of personal data of the plurality of first users;
requesting a validation response to a validation query from each of the first users, wherein the validation response is requested when each of the first users is enrolling or participating in a second stage of the multi-stage contest, wherein eligibility for a prize of the multi-stage contest for each of the first users requires the corresponding first user providing the validation response to the validation query, wherein the validation query includes content corresponding to that included in the second initial query such that the validation response of each of the first users can be compared to the predicted response of each of the respective first users to determine whether the initial predictive model correctly predicted the predicted response of each of the first users to the second initial query;
storing the validation response of each of the responding first users as validation response data, the validation response data forming a subset of the personal data set of each of the responding first users;
generating a validation predictive model during validation training of the machine learning program, wherein a validation training data set utilized during the validation training of the machine learning program comprises the personal data set of each of the plurality of users following the storing of the validation response of each of the responding first users as validation response data, wherein a difference between the predicted response and the validation response with respect to each of the first users is utilized as an error signal for correcting the initial predictive model during the generating of the validation predictive model;
deploying the trained validation predictive model; and
automatically selecting, by the computer, one of the account settings of;
(i) one of the plurality of first users to correspond to the predicted second response based on the initial predictive model; and
(ii) one of the third users to correspond to a predicted response of the one of the third users to one of the first query or the second query based on a prediction of the deployed validation predictive model, the automatic selecting of the one of the account settings of the one of the third users resulting in a change to a visually perceptible aspect of a graphical user interface of the software application and/or the website managed by the computer of the business entity when accessed by the one of the third users.
|