Credit card fraud detection model updating method and device based on joint learning

A detection model and model update technology, applied in the credit card field, can solve problems such as inability to protect data privacy, and achieve the effect of improving model efficiency

Pending Publication Date: 2020-06-23
SHENZHEN INST OF ADVANCED TECH
View PDF0 Cites 14 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The embodiment of the present invention provides a credit card fraud detection model update method and device based on federated learning to at least solve the technical problem that the existing fraud detection system cannot protect data privacy

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Credit card fraud detection model updating method and device based on joint learning
  • Credit card fraud detection model updating method and device based on joint learning
  • Credit card fraud detection model updating method and device based on joint learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0055] According to an embodiment of the present invention, a method for updating a credit card fraud detection model based on joint learning is provided, see figure 1 , including the following steps:

[0056] S101: Multiple clients use a convolutional neural network to construct a fraud detection model;

[0057] S102: Multiple clients input the local credit card transaction data set into the local fraud detection model, and calculate respective model update parameters;

[0058] S103: Perform parameter fusion calculation on the respective model update parameters of multiple clients to obtain new model parameters;

[0059] S104: Use the new model parameters to update the parameters of multiple local fraud detection models.

[0060] The present invention is different from traditional machine learning methods. Although the model used at the bottom layer is a convolutional neural network, the present invention is based on a joint learning algorithm and no longer needs to concent...

Embodiment 2

[0085] According to another embodiment of the present invention, a device is provided, see Figure 6 ,include:

[0086] A model construction unit 201, configured for multiple clients to construct a fraud detection model using a convolutional neural network;

[0087] The model update parameter unit 202 is used for multiple clients to input the local credit card transaction data set into the local fraud detection model, and calculate their respective model update parameters;

[0088] A parameter fusion calculation unit 203, configured to perform parameter fusion calculation on the respective model update parameters of multiple clients to obtain new model parameters;

[0089] The parameter updating unit 204 is configured to use new model parameters to update the parameters of multiple local fraud detection models.

[0090] While reducing the unbalanced influence of positive and negative samples, the device allows each client to obtain the knowledge learned by other clients by s...

Embodiment 3

[0096] A storage medium stores program files capable of implementing any one of the above joint learning-based credit card fraud detection model updating methods.

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to the field of credit cards, and particularly relates to a credit card fraud detection model updating method and device based on joint learning. The method comprises the following steps: firstly, constructing a fraud detection model by using a convolutional neural network; inputting the local credit card transaction data set into a local fraud detection model; calculating respective model updating parameters; and performing parameter fusion calculation on the plurality of model updating parameters to obtain new model parameters, performing parameter updating on a plurality of local fraud detection models by using the new model parameters, and training and updating own fraud detection models by means of data information of other clients in a form of sharing the modelparameters by different clients. And the model efficiency is improved on the premise of protecting the data privacy of each client from being invaded.

Description

technical field [0001] The present invention relates to the field of credit cards, in particular to a method and device for updating a credit card fraud detection model based on joint learning. Background technique [0002] With the rapid development of economic globalization in recent decades, credit cards have become more and more popular in business transactions. Correspondingly, the problem of credit card fraud has also emerged, and detecting fraudulent credit card transactions has become one of the challenges facing the banking industry. With the popularization of Internet technology, the traditional financial field has ushered in innovation. The credit card fraud detection system is an important research field of Internet financial technology. The financial platform calculates the risk value of each user and predicts whether the user is a fraudulent user, thereby helping banks or financial companies reduce risks and increase profits. The core algorithms of the system ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06Q40/02G06K9/62G06N3/04
CPCG06N3/045G06Q40/03G06F18/214
Inventor 阳文斯张昱航栗力须成忠
Owner SHENZHEN INST OF ADVANCED TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products