US 11,657,431 B1
Method, system, and apparatus for programmatically determining and adjusting electronic bid values for a digital content object
Rahul Anand, Sunnyvale, CA (US); Sandeep Dey, Bellevue, WA (US); Pravin P. Thampi, Kirkland, WA (US); and Zhi Guo, Sammamish, WA (US)
Assigned to GROUPON, INC., Chicago, IL (US)
Filed by Groupon, Inc., Chicago, IL (US)
Filed on Dec. 20, 2018, as Appl. No. 16/228,519.
Claims priority of provisional application 62/608,421, filed on Dec. 20, 2017.
Int. Cl. G06Q 30/00 (2023.01); G06Q 30/0273 (2023.01); G06N 20/00 (2019.01); H04L 67/01 (2022.01); G06Q 30/0251 (2023.01)
CPC G06Q 30/0275 (2013.01) [G06N 20/00 (2019.01); H04L 67/01 (2022.05); G06Q 30/0256 (2013.01)] 21 Claims
OG exemplary drawing
 
1. A method for adjusting an electronic bid value for a digital content object based on one or more network time period segments within one or more network time periods, wherein each of the one or more network time periods is divided equally by the one or more network time period segments, the method comprising:
generating one or more transaction signals by one or more client devices indicating the one or more client devices have completed a transaction with one or more device rendered objects;
receiving, from the one or more client devices, the one or more transaction signals associated with the one or more device rendered objects, wherein the one or more device rendered objects are associated with the digital content object, wherein the digital content object is associated with the electronic bid value;
performing a transaction between a client device of the one or more client devices and at least one device rendered object of the one or more device rendered objects in response to a rendering of the one or more device rendered objects via an interface of the client device, wherein each transaction signal of the one or more transaction signals represents a completion of the transaction between the client device of the one or more client devices and the at least one device rendered object of the one or more device rendered objects;
programmatically generating a first set of cumulative transaction values, each cumulative transaction value of the first set of cumulative transaction values associated with a different one of the one or more network time period segments within a first network time period, wherein each cumulative transaction value of the first set of cumulative transaction values is programmatically generated based on a total number of the one or more transaction signals associated with the one or more device rendered objects received during each of the one or more network time period segments within the first network time period;
programmatically generating a second set of cumulative transaction values, each cumulative transaction value of the second set of cumulative transaction values associated with a different one of the one or more network time period segments within a second network time period, wherein each cumulative transaction value of the second set of cumulative transaction values is programmatically generated based on a total number of the one or more transaction signals associated with the one or more device rendered objects received during each of the one or more network time period segments within the second network time period;
programmatically generating an accrued cumulative transaction value for each of the one or more network time period segments based on cumulating the first set of cumulative transaction values for each of the one or more network time period segments within the first network time period and the second set of cumulative transaction values for each of the one or more network time period segments within the second network time period;
generating a machine learning model comprising a linear model;
iteratively training the machine learning model using one or more training datasets comprising historical trend data associated with the one or more device rendered objects, wherein the machine learning model is associated with an r-squared value and a p-value, and wherein training the machine learning model comprises iteratively adjusting one or more parameters of the machine learning model based on successive comparisons of results of the machine learning model against target results until the r-squared value satisfies a threshold r-squared value and the p-value satisfies a threshold p-value;
programmatically generating, using the machine learning model, a slope value based on the accrued cumulative transaction value for each of the one or more network time period segments, wherein the slope value is associated with a particular network time period segment and indicates a transaction trend of the digital content object during the particular network time period segment; and
programmatically adjusting the electronic bid value for the digital content object during the particular network time period segment based at least on the slope value, wherein programmatically adjusting the electronic bid value for the digital content object comprises:
in circumstances where the slope value is larger than a first threshold slope value:
programmatically generating a multiplier value for the particular network time period segment; and
increasing the electronic bid value for the digital content object during the particular network time period segment based at least on the multiplier value.