Resource transferring monitoring method and device

Patent metadata

Field	Value
Publication number	US-11526889-B2
Application number	US-202016911089-A
Country	US
Kind code	B2
Filing date	Jun 24, 2020
Priority date	Feb 12, 2018
Publication date	Dec 13, 2022
Grant date	Dec 13, 2022

Abstract

Methods and apparatus, including computer programs encoded on computer storage media, for monitoring resource transfer are provided. One of the methods includes: by a server, receiving a resource deposit request from a resource deposit initiator; performing, using a first risk identification model, a first risk identification on the target account according to the resource deposit request to obtain a first risk identification result; receiving a resource withdrawal request from a resource withdrawal initiator, and the resource withdrawal request requesting a resource withdrawal from the target account to the recipient account; performing, using a second risk identification model, a second risk identification on the target account according to the resource withdrawal request to obtain a second risk identification result; and determining, using a third risk identification model, a resource transfer risk monitoring result of the target account according to the first risk identification result and the second risk identification result.

First claim

What is claimed is:

1. A method for monitoring a resource transaction, comprising:
training, by a processor, a first model with deposit risk training samples consisting of normal deposit transaction samples and fraud deposit transaction samples to obtain a neural network model of resource deposits and to output first risk identification results of the resource deposits from the neural network model, training a second model with withdrawal risk training samples consisting of normal withdrawal transaction samples and disposal transaction samples to obtain a gradient boosting regression model of resource withdrawals and to output second risk identification results of the resource withdrawals from the gradient boosting regression model, and training a third model with samples of the first risk identification results and samples of the second risk identification results to obtain a classification and regression tree model of resource disposals, wherein the deposit risk training samples are different from the withdrawal risk training samples;
receiving, by the processor, a resource deposit request for a resource deposit from a terminal of a resource deposit initiator, and the resource deposit request comprising an identifier of the resource deposit initiator and an identifier of a target account in which a resource is deposited;
performing, by the processor using the neural network model, a first risk identification on the target account according to the resource deposit request by at least inputting first association information related to the resource deposit request to the neural network model and outputting a first risk identification result of the resource deposit from the neural network model;
determining, by the processor, whether the resource deposit request is a fraudulent deposit request based on the first risk identification result of the resource deposit request;
in response to determining that the resource deposit request is not a fraudulent deposit request, obtaining, by the processor, first risk identification results of resource deposits being previously deposited into the target account, wherein the neural network model outputs the first risk identification results of the resource deposits being previously deposited into the target account;
receiving, by the processor, a resource withdrawal request for a resource withdrawal from a terminal of a resource withdrawal initiator, the resource withdrawal request comprising an identifier of the resource withdrawal initiator and an identifier of a recipient account, the resource withdrawal request requesting a resource withdrawal from the target account to the recipient account;
performing, by the processor using the gradient boosting regression model, a second risk identification on the target account according to the resource withdrawal request by at least inputting second association information related to the resource withdrawal request to the gradient boosting regression model and outputting a second risk identification result of the resource withdrawal from the gradient boosting regression model;
determining, by the processor, whether the resource withdrawal request is a fraudulent withdrawal request based on the second risk identification result;
in response to determining that the resource withdrawal request is not a fraudulent withdrawal request, obtaining, by the processor, the first risk identification results of resource deposits being previously deposited into the target account;
inputting, by the processor, the first risk identification results of the resource deposits being previously deposited into the target account and/or the second risk identification result to the classification and regression tree model;
determining, by the processor, at least one resource transfer risk identification strategy according to the first risk identification results of the resource deposits being previously deposited into the target account and/or the second risk identification result;
generating, by the processor, a determination result of whether the at least one resource transfer risk identification strategy meets a condition;
outputting, by the processor, a resource transaction risk monitoring result of the target account from the classification and regression tree model according to the determination result; and
determining, by the processor, whether to process the resource transaction comprising the resource deposit request and/or the resource withdrawal request according to the resource transaction risk monitoring result outputted from the classification and regression tree model.

2. The method according to claim 1 , wherein the performing a first risk identification on the target account according to the resource deposit request to obtain a first risk identification result comprises:
acquiring the first association information related to the resource deposit request, wherein the first association information comprises at least one of initiating account information, target account information including terminal activities and/or terminal environment, and first resource transaction information; and
performing the first risk identification on the target account according to the first association information using the neural network model to obtain the first risk identification result.

3. The method according to claim 1 ,
wherein the fraudulent deposit request indicates that the resource deposit request has a fraud risk; and
wherein the method further comprises, in response to determining that the resource deposit request is a fraudulent deposit request,
determining, by the processor, a deposit management and control mode corresponding to at least a threshold of the first risk identification result; and
triggering and executing, by the processor, the deposit management and control mode for not fulfilling the resource deposit request.

4. The method according to claim 1 , wherein the performing a second risk identification on the target account according to the resource withdrawal request to obtain a second risk identification result comprises:
acquiring the second association information related to the resource withdrawal request, wherein the second association information comprises at least one of target account information, second resource transaction information, or recipient account information; and
performing the second risk identification on the target account according to the second association information using the gradient boosting regression model to obtain the second risk identification result.

5. The method according to claim 1 , further comprising:
determining that the target account is a fraudulent account in response to that the at least one resource transfer risk identification strategy meets the condition, wherein the fraudulent account indicates that the target account has a fraud risk.

6. The method according to claim 5 , further comprising:
in response to determining that the target account is a fraudulent account,
determining, by the processor, a disposal management and control mode of the target account according to the at least one resource transfer risk identification strategy that meets the condition; and
triggering and executing, by the processor, the disposal management and control mode to manage and control the target account.

7. The method according to claim 1 , wherein the second risk identification result comprises second risk identification results of resource withdrawals being previously withdrawn from the target account.

8. An apparatus for monitoring a resource transaction, comprising a processor and a non-transitory computer-readable storage medium storing instructions executable by the processor to cause the apparatus to perform operations com

Assignees

• Advanced New Technologies Co Ltd

Inventors

• Ji Xiaoxi
• Gao Licui
• Chen Lujia
• Wang Weiqiang

Classifications

• G06N3/02

  Neural networks · CPC title

• G06N20/20

  Ensemble learning · CPC title

• G06N3/08

  Learning methods · CPC title

• G06Q20/382 Primary

  insuring higher security of transaction · CPC title

• G06Q20/36 Primary

  using electronic wallets or electronic money safes · CPC title