Anomaly detection algorithm integration method and system based on algorithm diversity

The invention provides an anomaly detection algorithm integration method based on algorithm diversity, and the method comprises the following steps: S01, building a plurality of basic trainers throughemploying a plurality of anomaly detection algorithms, carrying out the prediction of a sample set,...

Full description

Saved in:
Bibliographic Details
Main Authors XU MING, YIN QIAN'AN, LIU SHENG, TAO JINGLONG, WANG QIFAN, YU XIANZHE, LIANG SHUYUN, ZHOU XIAOYONG, MA YING, WEI GUOFU
Format Patent
LanguageChinese
English
Published 15.05.2020
Subjects
CALCULATING
COMPUTING
COUNTING
ELECTRIC DIGITAL DATA PROCESSING
PHYSICS
Online AccessGet full text

Cover

More Information
Summary:The invention provides an anomaly detection algorithm integration method based on algorithm diversity, and the method comprises the following steps: S01, building a plurality of basic trainers throughemploying a plurality of anomaly detection algorithms, carrying out the prediction of a sample set, and processing a prediction result to generate a pseudo label; S02, for each basic trainer, calculating a correlation coefficient of a prediction result of the basic trainer and a pseudo tag; S03, classifying all the anomaly detection algorithms; S04, for each classification, selecting a TOPN algorithm with the highest correlation coefficient and higher than a set threshold, and establishing an algorithm combination; and S05, performing anomaly detection by using an algorithm combination, andoutputting an abnormal point. A diversity model integration idea with supervised learning is introduced into anomaly detection, anomaly detection algorithms are classified according to implementationmechanisms of the algorithms
Bibliography:Application Number: CN201911406458