Title:
ABNORMAL USER IDENTIFICATION METHOD
Document Type and Number:
WIPO Patent Application WO/2019/218927
Kind Code:
A1
Abstract:
An abnormal user identification method, comprising: acquiring user behavior data of multiple users; extracting multiple user feature values of user behavior data of each user in multiple pre-set user behavior dimensions; determining a user feature vector of each user according to the multiple user feature values of each user; performing clustering processing on user feature vectors of multiple users by means of a pre-set clustering algorithm, so as to obtain multiple user classes; according to the user feature vectors comprised in each of the user classes, determining a central vector of each of the user classes; acquiring a difference feature vector of each of the user classes, wherein the difference feature vector is a user feature vector in the user classes that has a value of distance from the central vector of the user class that is not in a pre-set distance value range; and determining a user represented by the difference feature vector as an abnormal user.
Inventors:
GU CHENGJIE (CN)
ZHANG LI (CN)
ZHANG LI (CN)
Application Number:
PCT/CN2019/086232
Publication Date:
November 21, 2019
Filing Date:
May 09, 2019
Export Citation:
Assignee:
NEW H3C SECURITY TECH CO LTD (CN)
International Classes:
H04L29/06
Domestic Patent References:
| WO2016054988A1 | 2016-04-14 |
Foreign References:
| CN104268481A | 2015-01-07 | |||
| CN106649517A | 2017-05-10 | |||
| CN104239324A | 2014-12-24 | |||
| CN105681089A | 2016-06-15 |
Other References:
See also references of EP 3771168A4
Attorney, Agent or Firm:
PATENTSINO IP FIRM (CN)
Download PDF: