US 11,928,723 B2
Systems and methods for facilitating online search based on offline transactions
Yu Wang, Jersey City, NJ (US); Dianhui Zhu, Fremont, CA (US); Ayush Parshotam Ruchandani, Sunnyvale, CA (US); Cun Mu, Jersey City, NJ (US); Ying Michelle Sun, Palo Alto, CA (US); and Saumya Agarwal, Milpitas, CA (US)
Assigned to Walmart Apollo, LLC, Bentonville, AR (US)
Filed by Walmart Apollo, LLC, Bentonville, AR (US)
Filed on Jul. 29, 2021, as Appl. No. 17/389,111.
Prior Publication US 2023/0035109 A1, Feb. 2, 2023
Int. Cl. G06Q 30/00 (2023.01); G06F 18/2113 (2023.01); G06F 18/214 (2023.01); G06Q 20/12 (2012.01); G06Q 30/0601 (2023.01)
CPC G06Q 30/0639 (2013.01) [G06F 18/2113 (2023.01); G06F 18/214 (2023.01); G06Q 20/12 (2013.01); G06Q 30/0603 (2013.01); G06Q 30/0643 (2013.01)] 17 Claims
OG exemplary drawing
 
1. A system, comprising:
a computing device configured to:
obtain, from a first database via a transceiver, historical search data comprising a plurality of historical search results each comprising a plurality of items, an associated query, and an associated user;
obtain, from a second database via the transceiver, in-store purchase data comprising a plurality of offline transactions, wherein each of the plurality of offline transactions is associated with the associated user of at least one of the historical search results, wherein the in-store purchase data comprise at least one offline transaction initiated by the associated user, wherein the at least one offline transaction includes a transaction including at least one item of the plurality of items associated with the at least one of the historical search results, and wherein the transaction is within a predetermined time period of the associated query;
generate, by at least one processor and for each of the plurality of offline transactions, an order feedback signal by associating the in-store purchase data of the associated user with the plurality of items for the associated query; and
train, by the at least one processor a search model utilizing a training dataset comprising a plurality of training features based on the historical search results and the order feedback signal for each of the plurality of offline transactions;
receive, at a web server, a query; and
implement, by the at least one processor, the search model to generate an online search result.