US 12,299,700 B2
System and method for machine learning based generation of media structured products
Matthew C. Radcliffe, Brighton, MA (US)
Assigned to Isoquant Media, LLC, Boston, MA (US)
Filed by Isoquant Media, LLC, Brighton, MA (US)
Filed on May 7, 2018, as Appl. No. 15/972,642.
Claims priority of provisional application 62/507,972, filed on May 18, 2017.
Prior Publication US 2018/0336577 A1, Nov. 22, 2018
Int. Cl. G06Q 30/0201 (2023.01); G06N 20/00 (2019.01); G06N 20/10 (2019.01); G06N 3/02 (2006.01); G06N 5/01 (2023.01); G06N 5/025 (2023.01); G06N 7/01 (2023.01)
CPC G06Q 30/0201 (2013.01) [G06N 20/00 (2019.01); G06N 20/10 (2019.01); G06N 3/02 (2013.01); G06N 5/01 (2023.01); G06N 5/025 (2013.01); G06N 7/01 (2023.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
identifying, by a computing device, event specific data of a past media program and event specific data of a future media program, wherein the event specific data of the future media program includes social media activity for the future media program;
predicting a size of delivery of the future media program based upon, at least in part, the event specific data of the past media program and the social media activity for the future media program obtained via web scraping executed using a scripting language, wherein the predicted size of delivery of the future media program is a number of viewers predicted to view the future media program, wherein predicting the size of delivery of the future media program includes:
generating one or more quantitative time-series models, wherein the one or more quantitative time-series models includes a machine learning model;
training the one or more quantitative time-series models using the event specific data layered onto a base model to classify the base model according to the event specific data; and
iteratively back-testing the one or more quantitative time-series models using the event specific data of the past media program and the social media activity for the future media program pulled at different intervals as input to generate a best model fit;
generating a financial media product based upon, at least in part, the predicted size of delivery of the future media program, wherein generating the financial media product includes combining a plurality of financial media products and tranching the plurality of financial media products according to expected values on each tranch, wherein the expected values on the each tranch is determined based upon, at least in part, the iterative back-testing of the one or more quantitative time-series models using the event specific data of the past media program and the social media activity for the future media program;
issuing the financial media product through a bankruptcy-remote vehicle and passing funds from a sale of tranch notes of plurality of media products to a first business entity;
identifying an actual size of delivery, wherein the actual size of delivery is a number of viewers that actually viewed the future media program;
executing a first action if the number of viewers that actually viewed the future media program is greater than the number of viewers predicted to view the future media program based upon, at least in part, the financial media product, wherein the first action includes transferring a financial asset from a first business entity to a second business entity if the number of viewers that actually viewed the future media program is greater than the number of viewers predicted to view the future media program, wherein an amount of the financial asset for the first action is determined based upon, at least in part, the best model fit generated from the back-testing of the one or more quantitative time-series models using the event specific data pulled at different intervals as the input; and
executing a second action if the number of viewers that actually viewed the future media program is less than the number of viewers predicted to view the future media program based upon, at least in part, the financial media product, wherein the second action includes transferring a financial asset from the first business entity to a third business entity if the number of viewers that actually viewed the future media program is less than the number of viewers predicted to view the future media program, wherein an amount of the financial asset for the second action is determined based upon, at least in part, the best model fit generated from the back-testing of the one or more quantitative time-series models using the event specific data pulled at different intervals as the input, wherein the second business entity and the third business entity are different business entities.