Predicting future user transactions

Patent metadata

Field	Value
Publication number	US-11216832-B2
Application number	US-202016808002-A
Country	US
Kind code	B2
Filing date	Mar 3, 2020
Priority date	Jun 24, 2019
Publication date	Jan 4, 2022
Grant date	Jan 4, 2022

Abstract

A computer-implemented method includes: obtaining first environmental factor data of a first time period and first historical payment data of a target Internet of Things (IoT) device in the first time period, in which the first time period is before a first time point; obtaining second environmental factor data of a second time period that is after the first time point; and obtaining a prediction, using a neural network model, operating on at least one computing device, of an estimated statistical distribution of one or more users who execute transactions using the target IoT device in the second time period, in which the first environmental factor data, the first historical payment data, and the second environment factor data are inputs to the neural network model.

First claim

What is claimed is:

1. A computer-implemented method, comprising:
obtaining first environmental factor data of a first time period and first historical payment data of a target Internet of Things (IoT) device in the first time period, wherein the first time period is before a first time point, and
wherein the first historical payment data comprises data of payments made by offline user interaction using the target IoT device, and wherein the first environmental factor data comprises data of a context of the payments made by offline user interaction using the target IoT device;
obtaining second environmental factor data of a second time period that is after the first time point; and
obtaining a prediction, using a neural network model operating on at least one computing device, of an estimated statistical distribution of a plurality of other users who will execute transactions by offline user interaction using the target IoT device in the second time period,
wherein the first environmental factor data, the first historical payment data, and the second environmental factor data are inputs to the neural network model.

2. The computer-implemented method of claim 1 , comprising training the neural network model, wherein training the neural network model comprises:
obtaining third environmental factor data of a third time period and second historical payment data of the target IoT device in the third time period, wherein the third time period is before the first time point;
obtaining fourth environmental factor data of a fourth time period, wherein the fourth time period is before the first time point and after the third time period;
obtaining an actual statistical distribution of users who execute transactions by offline user interaction using the target IoT device in the fourth time period; and
training the neural network model by using the third environmental factor data, the second historical payment data, and the fourth environmental factor data as sample inputs to the neural network model, wherein the actual statistical distribution of users is a sample label of the sample inputs for the neural network model.

3. The computer-implemented method of claim 1 , wherein the first historical payment data comprises at least one of the following:
user profile data, merchant feature data, geographic location information, or sequence information.

4. The computer-implemented method of claim 1 , wherein the estimated statistical distribution is defined with respect to at least one of the following:
gender, age, or payment frequency.

5. The computer-implemented method of claim 1 , wherein the first historical payment data comprises user profile data, and wherein the method further comprises deriving the user profile data, wherein deriving the user profile data comprises:
obtaining raw profile data of each user of a plurality of users who interact offline with the target IoT device, wherein the raw profile data comprises at least one target feature, and wherein the at least one target feature comprises at least one of an age of the user, a preference label for the user, or a quantity of historical payments made by the user;
determining a first moment, a second moment, a third moment, and a summation result corresponding to each target feature of each user; and
using the first moment, the second moment, the third moment, and the summation result corresponding to each target feature of each user as the user profile data of the user.

6. The computer-implemented method of claim 1 , wherein the first environmental factor data comprises at least one of: a weather feature, a holiday feature, or a weekend and working day feature.

7. The computer-implemented method of claim 1 , wherein the first historical payment data comprises geographic location information, and further comprising determining the geographic location information, wherein determining the geographic location information comprises:
obtaining reported active locations of each user of a plurality of users who interact offline with the target IoT device, and a location of a merchant;
calculating, based on the obtained reported active locations and the location of the merchant, a distance between each user and the merchant; and
determining the geographical location information based on the distances between each user and the merchant.

8. The computer-implemented method of claim 1 , comprising:
subsequent to obtaining the prediction, emitting, by the target IoT device, an audio notification, the audio notification being targeted based on the estimated statistical distribution of the plurality of other users who will execute transactions by offline user interaction using the target IoT device in the second time period according to the prediction.

9. A non-transitory, computer-readable medium storing one or more instructions that, when executed by a computer system, cause the computer system to perform operations comprising:
obtaining first environmental factor data of a first time period and first historical payment data of a target Internet of Things (IoT) device in the first time period, wherein the first time period is before a first time point, and
wherein the first historical payment data comprises data of payments made by offline user interaction using the target IoT device, and wherein the first environmental factor data comprises data of a context of the payments made by offline user interaction using the target IoT device;
obtaining second environmental factor data of a second time period that is after the first time point; and
obtaining a prediction, using a neural network model operating on the computing system, of an estimated statistical distribution of a plurality of other users who will execute transactions by offline user interaction using the target IoT device in the second time period,
wherein the first environmental factor data, the first historical payment data, and the second environmental factor data are inputs to the neural network model.

10. The non-transitory, computer-readable medium of claim 9 , wherein the operations comprise training the neural network model, and wherein training the neural network model comprises:
obtaining third environmental factor data of a third time period and second historical payment data of the target IoT device in the third time period, wherein the third time period is before the first time point;
obtaining fourth environmental factor data of a fourth time period, wherein the fourth time period is before the first time point and after the third time period;
obtaining an actual statistical distribution of users who execute transactions by offline user interaction using the target IoT device in the fourth time period; and
training the neural network model by using the third environmental factor data, the second historical payment data, and the fourth environmental factor data as sample inputs to the neural network model, wherein the actual statistical distribution of users is a sample label of the sample inputs for the neural network model.

11. The non-transitory, computer-readable medium of claim 9 , wherein the first historical payment data comprises at least one of the following:
user profile data, merchant feature data, geographic location information, or sequence information.

12. The non-transitory, computer-readable medium of claim 9 , wherein the estimated statistical distribution is defined with respect to at least one of the following:
gender, age, or payment frequency.

13. The non-transitory, computer-readable medium of claim 9 , wherein the first historical payment data comprises user profile data, wherein the operations

Assignees

• Advanced New Technologies Co Ltd

Inventors

• He Yong
• Zhang Liang
• Chen Tian
• Fan Yicong
• Lei Jinhua
• Liu Bei
• Mei Han
• Ma Guolai

Classifications

• G06N3/044

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

• G06N3/0442

  characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06Q30/0205 Primary

  based on location or geographical consideration · CPC title

• G06N3/08

  Learning methods · CPC title