Deep Fault Prediction with Flexible Weighted Mining Based Alarm Correlation Analysis of Communication Networks

Author(s):  
Jun Jia ◽  
Chunyan Feng ◽  
Tiankui Zhang ◽  
Hailun Xia ◽  
Jinling Li ◽  
...  
2012 ◽  
Vol 198-199 ◽  
pp. 1539-1544 ◽  
Author(s):  
Pan Liu ◽  
Xing Ming Li ◽  
Jian Wu

The alarm correlation analysis based on fuzzy association rules mining is the popular and cutting-edge field of the network fault diagnosis research. In the application environment of alarms in communication networks, a new algorithm of the fuzziness of alarms which is called FKMA (Fuzzy K-Means of Alarms algorithm) is proposed .During the process of fuzziness, there are two methods of sorting the center. Simulations are carried out to the comparison of the two methods. The fuzziness of alarms is effectively realized. And fuzzy association rules mining are achieved. The advantages and efficiency of FKMA are demonstrated by experiments.


2014 ◽  
Vol 556-562 ◽  
pp. 6191-6195
Author(s):  
Yong Wei Wang ◽  
Hui Fang Su ◽  
Wei Qiu

This paper proposes a correlation analysis method based on fuzzy rules and artificial immune. Firstly, we adopt the alarms selection algorithm based on a sliding time window to improve the efficiency of selected alarm. Secondly, the analysis method based on fuzzy correlation rules is used to associate the known patterns static and rapidly. Then, using a method based on immune evolution to improve and adaptive the antibody so as to achieve the dynamic, intelligent correlation of unknown model. The experimental results in LLDOS1.0 and LLDOS2.0 show that the new method gets better accuracy than typical correlation methods, which can ensure the efficiency of correlation analysis and the adaptability of the correlation method.


Sign in / Sign up

Export Citation Format

Share Document