US 12,141,843 B2
Method and apparatus for advertisement information error detection and correction
Krishna Yalamanchili, San Jose, CA (US); Abilash Amarthaluri, San Jose, CA (US); Vishal Parmar, Fremont, CA (US); and Cecilia Seto, San Francisco, CA (US)
Assigned to Walmart Apollo, LLC, Bentonville, AR (US)
Filed by Walmart Apollo, LLC, Bentonville, AR (US)
Filed on Apr. 9, 2021, as Appl. No. 17/226,675.
Application 17/226,675 is a continuation of application No. 16/204,721, filed on Nov. 29, 2018, granted, now 10,997,635.
Prior Publication US 2021/0248649 A1, Aug. 12, 2021
This patent is subject to a terminal disclaimer.
Int. Cl. G06Q 30/00 (2023.01); G06Q 30/0241 (2023.01)
CPC G06Q 30/0276 (2013.01) [G06Q 30/0277 (2013.01)] 14 Claims
OG exemplary drawing
 
1. A system comprising:
a computing device communicatively coupled to a database and a plurality of sources, wherein each of the plurality of sources comprise at least one server, and wherein the computing device is configured to:
determine an attribute confidence value for a first attribute of an item based on a first number of sources of the plurality of sources reporting a first attribute value for the item and on a second number of sources of the plurality of sources reporting any other attribute value for the item;
obtain, from the database, a first source weight for a first source of the first number of sources;
apply the first source weight to the attribute confidence value to generate an adjusted attribute confidence value;
determine an item confidence value for the item based, at least partially, on the adjusted attribute confidence value for the first attribute of the item;
store the item confidence value for the item in the database;
generate an attribute error signal indicating an error with the first attribute value based on the determined item confidence value;
adjust the attribute confidence value based on whether previously received attribute values from the first source are correct; and
determine the item confidence value for the item based, at least partially, on the adjusted attribute confidence value;
determine one of a ratio of a number of times the previously received attribute values from the first source are correct to a total number of the previously received attribute values from the first source or a ratio of the plurality of sources that have provided a same attribute value for each of the previously received attribute values; and
determine, for each of the previously received attribute values, that they are correct if the determined ratio is beyond a threshold value; and
transmit the attribute error signal to the first source of the first number of sources to adjust the first attribute value based on the indicated error.