Title:
METHOD FOR DETECTING ABNORMAL TRANSACTION NODE, AND DEVICE
Document Type and Number:
WIPO Patent Application WO/2020/220758
Kind Code:
A1
Abstract:
A method for detecting an abnormal transaction node and a device, used to solve the problem of low efficiency and low accuracy of abnormal transaction detection in the prior art. The method comprises: determining, according to a transaction log covering a monitoring time interval, transaction feature values between transaction nodes under N transaction dimensions, and allocating, with respect to a single arbitrary transaction dimension, and according to the transaction feature values between the transaction nodes, the transaction nodes to transaction subsets under the transaction dimension; calculating, with respect to a single arbitrary transaction node, a cluster feature value of the transaction node in each of the transaction subsets at least according to a strong association relationship of the transaction subset having the transaction node; clustering all of the transaction nodes by using an unsupervised clustering algorithm and according to cluster feature values of the transaction nodes; and determining an abnormal transaction node on the basis of a clustering result. The method employs an association relationship between transaction nodes and an unsupervised clustering algorithm to detect an abnormal transaction node, thereby enhancing efficiency and accuracy of abnormal transaction detection.
Inventors:
TANG TAO (CN)
LIN JIALE (CN)
ZHAO JINTAO (CN)
ZHENG JIANBIN (CN)
LIU HONGBAO (CN)
LIN JIALE (CN)
ZHAO JINTAO (CN)
ZHENG JIANBIN (CN)
LIU HONGBAO (CN)
Application Number:
PCT/CN2020/071837
Publication Date:
November 05, 2020
Filing Date:
January 13, 2020
Export Citation:
Assignee:
CHINA UNIONPAY CO LTD (CN)
International Classes:
G06Q20/40; G06F17/00
Foreign References:
| CN106327340A | 2017-01-11 | |||
| CN109684673A | 2019-04-26 | |||
| CN110111113A | 2019-08-09 |
Attorney, Agent or Firm:
TDIP & PARTNERS (CN)
Download PDF: