量子近邻传播聚类算法的研究  

Study of quantum affinity propagation clustering algorithm

作　　者：苏一丹[1] 房骁 覃华[1] 王保锋 陈琴[1] SU Yi-dan;FANG Xiao;QIN Hua;WANG Bao-feng;CHEN Qin(School of Computer and Electronic Information,Guangxi University,Nanning 530004,China)

机构地区：[1]广西大学计算机与电子信息学院,广西南宁530004

出　　处：《广西大学学报（自然科学版）》2018年第2期561-568,共8页Journal of Guangxi University（Natural Science Edition）

基　　金：国家自然科学基金资助项目(61363027)

摘　　要：传统近邻传播聚类算法(affinity propagation,AP)的偏向参数与数据集相关,若不根据数据集作优化,将导致算法聚类精度低。为此提出一种量子近邻传播聚类算法,首先用量子叠加态编码偏向参数,再用量子旋转门搜索量子叠加态,找出近优的偏向参数供AP算法聚类,达到自适应优化AP偏向参数的目的。在UCI数据集上的实验结果表明,本文提出的算法聚类精度比同类算法更优,计算时间和迭代次数亦优于传统AP算法,引入量子计算优化近邻传播聚类算法的思路是可行的。The preference of the affinity propagation(AP)clustering algorithm is related to the data sets.If the algorithm is not optimized according to the data sets,the clustering accuracy of the algorithm will be reduced.A quantum affinity propagation clustering algorithm is proposed in this paper.Firstly,the preference parameter is encoded by the quantum superposition state.Secondly,the superposition state of the quantum is searched by the quantum rotate gate,with the purpose of finding out the approximately optimal preference value for AP algorithm clustering.By that method,the AP preference parameter can be optimized adaptively.Experimental results on the UCI data sets show that the proposed algorithm gets better clustering accuracy against several other similar algorithms,and the computation time and the number of iterations are superior to the traditional AP algorithm.Therefore,the idea that quantum computation is introduced in to optimize the affinity propagation clustering algorithm is feasible.

关 键 词：近邻传播聚类 偏向参数优化 量子计算 量子旋转门 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]