Study on a New Clustering Optimization Algorithm and Its Application in Network Faults Analysis
Peiqi Liu, LI Zeng-zhi · 2009
The iterative optimization algorithm is a traditional classification method of the pattern recognition. In the iterative optimization algorithm, the primary center of classes is selected by random method. This choice method causes the iterative time increase greatly in the optimization at anaphase. It also has some serious defects which are the selected samples blindly, the presented a local extremum and the omitted clustering tendency of samples. By the researching and analyzing the iterative optimization algorithm, the newly algorithm, clustering optimization algorithm based on neighborhood of samples distribution, is designed according to the conception of the clustering tendency and neighborhood of patterns in this paper. The time complexity of the newly algorithm is O(n) and n is a number of samples in sets. This algorithm is applied in the faults analysis in network management based on SNMP protocol. The analysis results were consistent with faults type and this algorithm provides a feasible method for faults analysis.