A differential clustering algorithm based on elite strategy
/ 0
- Details
- Category: IT technologies
- Last Updated on Thursday, 23 June 2016 21:54
- Published on Thursday, 23 June 2016 21:54
- Hits: 4214
Authors:
Xiongjun Wen, Hunan International Economics University, Changsha, Hunan, China
Qun Zhou, Hunan International Economics University, Changsha, Hunan, China
Sheng Huang, Hunan International Economics University, Changsha, Hunan, China
Abstract:
Purpose. Cluster analysis is not only an important research field of data mining but also a significant means and method in data partitioning or packet processing. The research aims to further improve the effect of clustering algorithm and overcome the existing defects of differential evolution (DE). The research achievements are intended to be used in the cluster analysis to obtain better clustering effect.
Methodology. We have made in-depth research with regards to DE algorithm and cluster analysis, discussed the effect of K-means as well as the flowchart and computing method of the fitness function. The influence of different differential operations plays on the performance have been analyzed.
Findings. Firstly, we explained the fundamental ideas and methods of cluster analysis and DE algorithm. Then we illustrated how the improved DE algorithm realizes the cluster analysis. Finally, we conducted simulation experiment of cluster analysis on four artificial data through the clustering algorithm based on elite strategy DE algorithm so that we can verify the feasibility and validity of the new method.
Originality. We developed an elite strategy DE algorithm and used it in K-means cluster analysis. Since DE algorithm is a method to search the optimal solution by simulating natural evolution process, its outstanding features are its implicit parallelism and ability to utilize the global information effectively, therefore, the new and improved algorithm has stronger robustness and it can avoid getting trapped in a local optimum and greatly enhances the clustering effect. The research on this aspect has not been found at present.
Practical value. By applying elite strategy DE algorithm in K-means cluster analysis, we can improve the efficiency and accuracy of cluster analysis. The result of the simulation experiment showed that the new method presented in this paper has significantly improved the optimization performance, which verifies the feasibility and effectiveness of this new method.
Список літератури / References
1. Enmei Tu, Longbing Cao, Jie Yang and Nicola Kasabov, 2014. A novel graph-based K-means for nonlinear manifold clustering and representative selection. Neurocomputing, vol.143, no.2, pp. 109‒122.
2. Michio Yamamoto and Yoshikazu Terada, 2014. Functional factorial image-means analysis. Computational Statistics & Data Analysis, vol.79, no.11, pp. 133‒148.
3. Basu, M., 2014. Improved differential evolution for economic dispatch. International Journal of Electrical Power & Energy Systems, vol.63, no.12, pp. 855‒861.
4. Ali Wagdy Mohamed, 2014. RDEL: Restart differential evolution algorithm with local search mutation for global numerical optimization. Egyptian Informatics Journal, vol.15, no.3, pp. 175‒188.
5. Pratyay Kuila and Prasanta K. Jana, 2014. A novel differential evolution based clustering algorithm for wireless sensor networks. Applied Soft Computing, vol.25, no.12, pp. 414‒425.
6. Das, S., Konar, A., Chakraborty, U.K., 2005. Improved differential evolution algorithms for handling noisy optimization problems. In: IEEE. The 2005 IEEE Congress on Evolutionary Computation, vol.2, pp. 1691‒1698.
7. Qingya Zhou, 2014. The research of differential evolution under dynamic environment. Zhengzhou University, China.
8. Md Anisur Rahman and Md Zahidul Islam, 2014. A hybrid clustering technique combining a novel genetic algorithm with K-Means. Knowledge-Based Systems, vol.71, no.11, pp. 345‒365.
9. Grigorios Tzortzis and Aristidis Likas, 2014. The MinMax K-Means clustering algorithm. Pattern Recognition, vol.47, no.7, pp. 2505‒2516.
10. Velmurugan, T., 2014. Performance based analysis between K-means and fuzzy C-means clustering algorithms for connection oriented telecommunication data. Applied Soft Computing, vol.19, no.6, pp. 134‒146.
 2016_02_Xiongjun
|
|
 2016-06-21  1.27 MB  734 |
|
Archive