Systems and methods for prediction of item quantity

Patent metadata

Field	Value
Publication number	US-11544763-B2
Application number	US-202016778905-A
Country	US
Kind code	B2
Filing date	Jan 31, 2020
Priority date	Jan 31, 2020
Publication date	Jan 3, 2023
Grant date	Jan 3, 2023

Abstract

Systems and methods including one or more processors and one or more non-transitory storage devices storing computing instructions configured to run on the one or more processors and perform acts of receiving a user identifier, receiving an item identifier, determining user item quantity information related to quantities of the item previously selected by the user, determining a respective household size for each user, and determining aggregate household item quantity information related to quantities of the item previously selected by an aggregate of users of the same household size. If a first threshold level of the quantity of transactions is met, a recommended quantity is based on the user item quantity information, and if not, the recommended quantity is based on the aggregate household item quantity information. The user interface of the electronic device is updated to notify the user of the recommended quantity. Other embodiments are disclosed herein.

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, that when executed on the one or more processors, cause the one or more processors to perform:
receiving a user identifier from a user interface of an electronic device, wherein the user identifier identifies a user from among a set of users;
receiving an item identifier associated with the user identifier, wherein the item identifier identifies an item in a catalog comprising a set of items;
receiving product-type information that identifies one or more products of a type related to the item;
determining user item quantity information related to quantities of the item previously selected by the user in first prior user transactions by the user involving the item in a predetermined time period;
determining a respective household size for each user in the set of users;
determining aggregate household item quantity information related to quantities of the item previously selected by each user in a portion of the set of users in first prior item transactions by the portion of the set of users involving the item in the predetermined time period, wherein each user in the portion of the set of users has a household size corresponding to the respective household size of the user;
generating, using a combination of two learning models comprising a generator model and a discriminator model, a predicted quantity of the item from latent data, wherein the latent data is used as input into the generator model to output generated data to learn a distribution of real data, and wherein the generated data that is output from the generator model and the real data are used as input into the discriminator model to estimate a distribution of the latent data to a real dataset;
using a respective loss function that specifies a generator loss for the generated data that is output from the generator model and a discriminator loss for an output of the discriminator model;
propagating the combination of the generator loss and the discriminator loss back through the generator model until the discriminator model can no longer distinguish the generated data from the real data in the real dataset;
determining whether a quantity of the first prior user transactions by the user meets or exceeds a first threshold level;
when the quantity of the first prior user transactions is determined to meet or exceed the first threshold level, determining a recommended quantity of the item for notification to the user based on the user item quantity information and updating the user interface of the electronic device to notify the user of the recommended quantity of the item;
when the quantity of the first prior user transactions is determined to be greater than zero but does not meet or exceed the first threshold level, determining the recommended quantity of the item for notification to the user based on the aggregate household item quantity information and updating the user interface of the electronic device to notify the user of the recommended quantity of the item;
when the quantity of the first prior user transactions is determined to be zero, determining user product type quantity information related to quantities of the one or more products of the type related to the item previously selected by the user in second prior user transactions by the user involving the one or more products of the type related to the item; and
determining an aggregate household product type quantity information related to quantities of the one or more products of the type related to the item previously selected by each user in the portion of the set of users in second prior item transactions by the portion of the set of users involving the one or more products of the type related to the item.

2. The system of claim 1 , wherein determining the recommended quantity of the item for notification to the user based on the user item quantity information comprises:
selecting, as the recommended quantity, either a most frequently selected quantity of the item selected by the user, or a median selected quantity of the item selected by the user,
wherein:
the generator model comprises a long short term memory recurrent neural network; and
the discriminator model comprises a convolutional neural network.

3. The system of claim 1 , wherein determining the recommended quantity of the item for notification to the user based on the aggregate household item quantity information, comprises:
determining an item confidence interval for the aggregate household item quantity information for the item for the household size corresponding to the respective household size of the user;
determining whether a most frequently selected quantity of the item is within the item confidence interval for the item; and
when the most frequently selected quantity is determined to be within the item confidence interval for the item, selecting, as the recommended quantity, the most frequently selected quantity of the item.

4. The system of claim 1 , wherein the computing instructions, when executed on the one or more processors, further cause the one or more processors to perform:
determining whether a quantity of the second prior user transactions meets or exceeds a second threshold level;
when the quantity of the second prior user transactions is determined to meet or exceed the second threshold level, determining the recommended quantity of the item for notification to the user based on the user product type quantity information and updating the user interface of the electronic device to notify the user of the recommended quantity of the item; and
when the quantity of the second prior user transactions is determined to be greater than zero but does not meet or exceed the second threshold level, determining the recommended quantity of the item for notification to the user based on the aggregate household product type quantity information and updating the user interface of the electronic device to notify the user of the recommended quantity of the item.

5. The system of claim 4 , wherein determining the recommended quantity of the item for notification to the user based on the user product type quantity information comprises:
selecting, as the recommended quantity, either a most frequently selected quantity of the one or more products of the type related to the item selected by the user, or a median selected quantity of the one or more products of the type related to the item selected by the user.

6. The system of claim 4 , wherein determining the recommended quantity of the item for notification to the user based on the aggregate household product type quantity information comprises:
determining a product type confidence interval for the aggregate household product type quantity information for the one or more products of the type related to the item for the household size corresponding to the respective household size of the user.

7. The system of claim 4 , wherein the computing instructions, when executed on the one or more processors, further cause the one or more processors to perform:
determining whether a most frequently selected quantity of the one or more products of the type related to the item is within product type confidence interval for the one or more products of the type related to the item;
when the most frequently selected quantity is determined to be within the product type confidence interval for the one or more products of the type related to the item, selecting, as the recommended quantity, the most frequently selected quantity of the one or more products of the type related to the item; and
when the quantity of the first prior user transactions is determined to be zero an

Assignees

• Walmart Apollo Llc

Inventors

• Iyer Rahul Radhakrishnan
• Gupta Shubham
• Kanumala Praveenkumar
• Guo Stephen Dean
• Achan Kannan

Classifications

• G06Q30/0631 Primary

  Recommending goods or services · CPC title

• G06Q30/0633

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

• G06Q30/0603

  Catalogue creation or management · CPC title

• G06Q30/0641

  utilising user interfaces specially adapted for shopping · CPC title

• G06F3/0482

  Interaction with lists of selectable items, e.g. menus · CPC title