扩散涌现的增量聚类算法

摘要

为了有效聚类动态数据,妥善处理已存在的类簇与新增数据的关系,高效利用计算资源,提高聚类的效率,扩散涌现的增量聚类算法被提出.该算法在扩散涌现聚类算法的基础上,利用近邻传播算法完善了算法的分裂机制,实现了新旧数据的有效聚合.实验结果表明,该算法有效实现了动态数据的聚类,提高了聚合动态数据的效率和资源的利用率.%In order to effectively cluster dynamic data, properly handle the relationship between the new data and the existing class, and improve clustering efficiency and utilization of computing resources, a diffused and emerging incremental clustering algorithm (DEICA) is proposed. On the basis of diffused and emerging clustering algorithm (DECA), affinity propagation (AP) is used to improve division mechanism of the algorithm, so old and new data are efficiently clustered. Many experiments demonstrate our algorithm can achieve clustering of dynamic data and improve the efficiency of dynamic data aggregation and the utilization of resource.