Automatically determining in real-time a triggering model for personalized recommendations

Patent metadata

Field	Value
Publication number	US-11107144-B2
Application number	US-202016779541-A
Country	US
Kind code	B2
Filing date	Jan 31, 2020
Priority date	Jan 31, 2020
Publication date	Aug 31, 2021
Grant date	Aug 31, 2021

Abstract

A method including building a recommendation triggering model. The method can include receiving, via a user device of a user through a network, an add-to-cart command associated with an anchor item in a session by the user. The method further can include determining, in real-time after receiving the add-to-cart command, a recommendation for one or more complementary items based at least in part on: (a) the anchor item; and (b) a user profile of the user. The method also can include determining, in real-time after determining the recommendation, a recommendation confidence for the recommendation based at least in part on one or more of: (a) the user profile; (b) the anchor item; (c) the one or more complementary items; or (d) one or more feedbacks from the user associated with one or more prior recommendations in the session. The method additionally can include after determining the recommendation confidence, when the recommendation confidence is positive, transmitting, in real-time through the network, the one or more complementary items to be presented to the user via the user device. The method likewise can include after determining the recommendation confidence, when the recommendation confidence is not positive, refraining from transmitting the one or more complementary items to the user. Other embodiments are disclosed.

First claim

What is claimed is:

1. A system comprising:
one or more processors; and
one or more non-transitory computer-readable media storing computing instructions configured to run on the one or more processors and perform:
receiving, via a user device of a user through a network, an add-to-cart command associated with an anchor item in a session by the user;
determining, in real-time after receiving the add-to-cart command, a recommendation for one or more complementary items based at least in part on: (a) the anchor item; and (b) a user profile of the user;
determining, in real-time after determining the recommendation, a recommendation confidence for the recommendation based at least in part on one or more of:
the user profile;
the anchor item;
the one or more complementary items; or
one or more feedbacks from the user associated with one or more prior recommendations in the session; and

after determining the recommendation confidence, when the recommendation confidence is positive, transmitting, in real-time through the network, the one or more complementary items to be presented to the user via the user device; and
after determining the recommendation confidence, when the recommendation confidence is not positive, refraining from transmitting the one or more complementary items to the user.

2. The system in claim 1 , wherein the computing instructions are further configured to perform:
prior to determining the recommendation confidence for the recommendation, removing a duplicate item of the one or more complementary items when at least one of:
one or more prior declined recommendations comprise the duplicate item; or
a cart for the user comprises the duplicate item.

3. The system in claim 1 , wherein determining the recommendation confidence further comprises:
determining an approximate entropy for one or more joint probability distributions, each respective joint probability distribution of the one or more joint probability distributions being associated with the user, the anchor item, and each of the one or more complementary items;
after determining the approximate entropy, when the approximate entropy is less than a predetermined uncertainty threshold, the recommendation confidence is positive; and
after determining the approximate entropy, when the approximate entropy is not less than the predetermined uncertainty threshold, the recommendation confidence is negative.

4. The system in claim 3 , wherein:
the user profile comprises a transactional history; and
a respective joint probability distribution for the user, the anchor item, and each of the one or more complementary items is determined based at least in part on one or more of:
a recommendation add-to-cart score associated with one or more of: the transactional history of the user profile, the one or more prior recommendations, or the one or more feedbacks;
an add-to-cart score for the anchor item associated with the user; or
a respective likelihood score for each of the one or more complementary items associated with the anchor item and the user.

5. The system in claim 4 , wherein:
each of the one or more complementary items is associated with one or more respective attributes;
the user profile further comprises one or more user attribute preferences for one or more respective attributes associated with each of the one or more complementary items; and
the respective likelihood score for each of the one or more complementary items is determined based at least in part on the one or more user attribute preferences.

6. The system in claim 1 , wherein:
determining the recommendation further comprises:
determining each of the one or more complementary items based at least in part on one or more of:
one or more complementary item types for an item type of the anchor item;
a respective item-item complementarity score for the each of the one or more complementary items associated with the anchor item; or
a respective user-item compatibility score for the each of the one or more complementary items associated with one or more system users, the one or more system users comprising the user.

7. The system in claim 6 , wherein:
each of the one or more complementary items is associated with a respective co-bought score determined based at least in part on one of:
the respective item-item complementarity score;
the respective user-item compatibility score; or
a respective item-type compatibility score; and

determining the recommendation further comprises:
determining a respective rank for each of the one or more complementary items based on the respective co-bought score.

8. The system in claim 1 , wherein:
determining the recommendation further comprises re-ranking the one or more complementary items based on a respective likelihood score for each of the one or more complementary items associated with the anchor item and the user.

9. The system in claim 1 , wherein the computing instructions are further configured to perform:
prior to determining the recommendation confidence for the recommendation, re-ranking the one or more complementary items by boosting a boosted item of the one or more complementary items, when at least one of:
one or more favorites of the user profile comprises the boosted item; or
one or more promotional items of a retailer comprises the boosted item.

10. The system in claim 1 , wherein:
transmitting the one or more complementary items to be presented to the user via the user device is performed when one or more of:
at a check out by the user;
in real-time after determining the recommendation confidence for the recommendation; or
at a session-change by the system.

11. A method being implemented via execution of computing instructions configured to run at one or more processors and stored at one or more non-transitory computer-readable media, the method comprising:
receiving, via a user device of a user through a network, an add-to-cart command associated with an anchor item in a session by the user;
determining, in real-time after receiving the add-to-cart command, a recommendation for one or more complementary items based at least in part on: (a) the anchor item; and (b) a user profile of the user;
determining, in real-time after determining the recommendation, a recommendation confidence for the recommendation based at least in part on one or more of:
the user profile;
the anchor item;
the one or more complementary items; or
one or more feedbacks from the user associated with one or more prior recommendations in the session;

after determining the recommendation confidence, when the recommendation confidence is positive, transmitting, in real-time through the network, the one or more complementary items to be presented to the user via the user device; and
after determining the recommendation confidence, when the recommendation confidence is not positive, refraining from transmitting the one or more complementary items to the user.

12. The method in claim 11 further comprising:
prior to determining the recommendation confidence for the recommendation, removing a duplicate item of the one or more complementary items when at least one of:
one or more prior declined recommendations comprise the duplicate item; or
a cart for the user comprises the duplicate item.

13. The method in claim 11 , wherein determining the recommendation confidence further comprises:
determining an approximate entropy for one or more joint probability distributions, each respective joint probability distribution of the one or more joint probability distributions being associated with the user, the anchor item, and

Assignees

• Walmart Apollo Llc

Inventors

• Mantha Aditya
• Iyer Rahul Radhakrishnan
• Kedia Shashank
• Gupta Shubham
• Kanumala Praveenkumar
• Guo Stephen Dean
• Achan Kannan

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06N3/044

  Recurrent networks, e.g. Hopfield networks · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06N5/025

  Extracting rules from data · CPC title

• G06Q30/0633

  Managing shopping lists, e.g. compiling or processing purchase lists (shipping orders G06Q10/083; order filling G06Q10/087) · CPC title