摘要
已有的粒子群模糊聚类算法需要设置粒子群参数并且收敛速度较慢,对此提出一种基于改进粒子群与模糊c-means的模糊聚类算法。首先,使用模糊c-means算法生成一组起始解,提高粒子群演化的方向性;然后,使用改进的自适应粒子群优化方法对数据进行训练与优化,训练过程中自适应地调节粒子群参数;最终,采用模糊c-means算法进行模糊聚类过程。对比实验结果表明,所提方法大幅度提高了计算速度,并获得了较高的聚类性能。
The existing PSO fuzzy clustering algorithms need to set the PSO parameters and converge very slowly, a fuzzy c-means and adaptive PSO based fuzzy clustering algorithm was proposed for that problem. Firstly, the fuzzy c-means algorithm is used to generate the initial solution, leading to a more directed search process. Then, the improved adaptive PSO is used to train and optimize the dataset, and the PSO parameters are adjusted adaptively in the training process to achieve a better optimal result. Lastly, the fuzzy c-means algorithm is used for fuzzy clustering. Compared ex- periments results show that the proposed method improves computational speed greatly and achieve good clustering per- formance.
作者
耿宗科
王长宾
张振国
GENG Zong-ke WANG Chang-bin ZHANG Zhen-guo(College of Mathematics and Information Science, Hebei Normal University, Shijiazhuang 050024,Chin)
出处
《计算机科学》
CSCD
北大核心
2016年第8期267-272,共6页
Computer Science
基金
国家自然科学基金项目(71271067)资助
关键词
粒子群优化
参数调节
模糊聚类算法
自适应调节
收敛速度
Particle swarm optimization, Parameter adjustment, Fuzzy clustering algorithm, Aadaptively adjustment,Convergence speed