DocumentCode
437557
Title
A fuzzy rules based approach for performance anomaly detection
Author
Xu, Jian ; You, Jing ; Liu, Fengyu
Author_Institution
Dept. of Comput. Sci. & Technol., Nanjing Univ. of Aeronaut. & Astronaut., China
fYear
2005
fDate
19-22 March 2005
Firstpage
44
Lastpage
48
Abstract
This paper presents a new approach inspired by immunology for system performance anomaly detection, which combines the negative selection algorithm (NSA) and genetic algorithm, generating a set of fuzzy rules that can characterize the normal and the abnormal. NSA serves as a filter to eliminate invalid detectors and reduce search space. Experiments with synthetic and real data sets are performed to show the applicability of the proposed approach.
Keywords
fuzzy logic; genetic algorithms; performance evaluation; systems analysis; fuzzy rules; genetic algorithm; immunology; negative selection algorithm; performance anomaly detection; Character generation; Computer science; Detectors; Fault detection; Filters; Fuzzy sets; Fuzzy systems; Genetic algorithms; Space technology; System performance;
fLanguage
English
Publisher
ieee
Conference_Titel
Networking, Sensing and Control, 2005. Proceedings. 2005 IEEE
Print_ISBN
0-7803-8812-7
Type
conf
DOI
10.1109/ICNSC.2005.1461158
Filename
1461158
Link To Document