基于粒子群优化算法的数据流聚类算法

针对当前基于滑动窗口的聚类算法中对原始数据信息的损失问题和提高聚类质量和准确性，在现有基于滑动窗口模型数据流聚类算法的基础上，提出了一种基于群体协作的粒子群优化算法（PSO）的新数据流聚类算法。这种优化的新数据流聚类算法利用改进的时间聚类特征指数直方图作为数据流的概要结构以及应用PSO在聚类过程中对聚类质量的局部迭代优化。实验结果表明，此方法有效减少了内存的开销，解决了对原始数据信息损失的问题。与传统的数据流聚类算法相比，基于粒子群优化算法的数据流聚类算法在聚类质量和准确性上明显优于传统的数据流聚类算法。

Clustering Evolving Data Streams Based on Particle Swarm Optimization

In view of the current based on sliding windows clustering algorithm of original data information loss problem and improve the cluster quality and accuracy, in the existing basis for data flow clustering algorithm based on sliding window model, proposed based on group collaboration of particle swarm optimization algorithm （PSO） of new data flow clustering algorithm, the optimization of new data flow clustering algorithm by means of improved time clustering indexes as data flow histogram summary of structure and the cluster quality of local iterative optimization in clustering process using the PSO. The experiment results show that this method is effective to reduce the memory spending, and solved the problem of original data loss. Compared with the traditional data flow clustering algorithms, based on the particle swarm optimization algorithm of data flow clustering algorithm evidently excels the traditional flow of data clustering algorithm in cluster quality and accuracy.