Bank electronic channel abnormal transaction determination method based on semi-supervised learning

A semi-supervised learning, abnormal transaction technology, applied in the field of abnormal transaction determination in bank electronic channels, can solve problems such as poor accuracy, high difficulty, and low efficiency, and achieve the effect of reducing difficulty, improving work efficiency, and improving accuracy.

Pending Publication Date: 2019-07-23
HARBIN ENG UNIV
View PDF6 Cites 9 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to propose a semi-supervised learning-based method for determining abnormal transactions in bank electronic channels in view of the problems of high difficulty, low efficiency, and poor accuracy in the existing abnormal transaction determination technology

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
  • Bank electronic channel abnormal transaction determination method based on semi-supervised learning
  • Bank electronic channel abnormal transaction determination method based on semi-supervised learning
  • Bank electronic channel abnormal transaction determination method based on semi-supervised learning

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0033] Specific implementation mode one: the following combination figure 1 This embodiment is described. In this embodiment, a semi-supervised learning-based method for determining abnormal transactions in bank electronic channels includes the following steps:

[0034] Step 1: Input a labeled sample set L and an unlabeled sample set U, perform SMUC clustering on L and U, and select the first r unlabeled samples with a high degree of membership to construct a sample set R 1 , if there are less than r unlabeled samples with high membership degree, use these unlabeled samples directly to construct R 1 , where r is the number of unlabeled samples selected after SMUC clustering, and k is the nearest neighbor parameter;

[0035] Step 2: Use L to train the classifier h on two different views 1 and h 2 , use the trained classifier to R 1 Classification;

[0036] Step 3: Determine h 1 、h 2 to R 1 Whether the classification of each sample in is consistent, if the record is cons...

specific Embodiment approach 2

[0080] Embodiment 2: This embodiment is a further description of Embodiment 1. The difference between this embodiment and Embodiment 1 is that the specific steps of SMUC clustering in Step 1 are:

[0081] Step 11: Initialize the dataset;

[0082] Step 1 and 2: Initialize the membership array;

[0083] Step 13: update the cluster center according to the membership array;

[0084] Step 14: update the membership degree array according to the cluster center;

[0085] Step 15: Determine whether the end condition is met, that is, the center of mass does not change any more, if not, repeat steps 12 to 14.

specific Embodiment approach 3

[0086] Specific embodiment three: this embodiment is a further description of specific embodiment one, and the difference between this embodiment and specific embodiment one is the specific steps of the Baum-Welch algorithm in the described step six:

[0087] Step 61: Randomly set the initial model M 0 ;

[0088] Step 62: Based on M 0 And the sequence of observations O, train a new model M *;

[0089] Step 63: If the condition is satisfied||M * -M 0 ||<ε, that is, the training has reached the expected effect, and it is over. If it is not satisfied, go to step 64;

[0090] Step 64: Let M 0 = M * , continue to work in step 62.

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 discloses a bank electronic channel abnormal transaction determination method based on semi-supervised learning, and relates to the technical field of machine learning. The invention aims to solve problems of high difficulty, low efficiency and poor accuracy in the existing abnormal transaction determination technology. According to the method, further integration and optimization are carried out on the basis that a hidden Markov model and a time sequence model (ARIMA) establish an account-level historical transaction sequence model, and an abnormal transaction behavior is predicted by combining semi-supervised clustering learning on the basis of HMM. Transaction data of each time section are converted into a time sequence vector through semi-supervised clustering learning, and semi-supervised learning is utilized to overcome the problem that label data is rare, an HMM is utilized to fit transaction vectors of everyone to generate a corresponding model, and the semi-supervised learning and the HMM are combined to improve the accuracy of anomaly recognition from two aspects of cross section data and time sequence data. Machine learning is adopted to solve the problem of abnormal transaction determination. Compared with a traditional expert method, the difficulty is greatly reduced, and the working efficiency is improved.

Description

technical field [0001] The invention relates to the technical field of machine learning, in particular to a method for determining abnormal transactions in bank electronic channels. Background technique [0002] Regulatory agencies such as the People's Bank of China and the China Banking Regulatory Commission have always attached great importance to the security of electronic channels, and have put forward a number of clear and specific regulatory requirements. Among them, in the "Notice of the General Office of the China Banking Regulatory Commission on Further Strengthening the Risk Prevention and Control of Electronic Channels", the China Banking Regulatory Commission has made clear instructions for commercial banks to strengthen the research and system construction of electronic channel suspicious transaction monitoring mechanisms. "All banking and other financial institutions should further research and build and improve the electronic channel risk monitoring system, st...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/62G06Q40/04
CPCG06Q40/04G06F18/2321G06F18/2155G06F18/29
Inventor 董宇欣张艺薰费腾王红滨温启明孙孟昊
Owner HARBIN ENG UNIV
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

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap