摘要
针对密度峰值算法在社区划分应用中因截断距离的选取以及仅考虑社区网络拓扑结构而带来的不准确等问题,提出一种融合结点属性和网络拓扑结构的密度峰值社区检测算法。以用户网络拓扑结构计算用户结点的直接邻居与间接邻居的度表示结点的局部密度;以用户对商品的评论信息的主题属性为用户结点的属性,结合网络拓扑结构计算用户间的相似度,从而得到用户间的相对距离;选出关键点作为社区的中心结点,完成社区划分。实验表明:所提出的算法准确度与归一化信息指标均优于基线模型算法,提高了电商网络中社区检测算法准确度,实现了高效的社区划分。
In order to solve the problems such as the selection of cut-off distance and the inaccuracy of community partition,a community detection algorithm based on density peak algorithm is proposed,which combines node attributes and network topology.The algorithm first calculates the degree of the direct neighbor and the indirect neighbor of the user node to express the local density of the node,then takes the subject attribute of the user's comment information to the commodity as the attribute of the user node,and calculates the similarity degree between the users based on the network topology,so as to obtain the relative distance between the users,and finally selects the key point as the center node of the community click to complete the community division.Experimental results show that the algorithm is better than the baseline model algorithm in accuracy and normalized information index,which improves the accuracy of community detection algorithm in e-commerce network and realizes efficient community division.
作者
高娟
张晓滨
GAO Juan;ZHANG Xiaobin(School of Computer Science,Xi’an Polytechnic University, Xi’an 710048,China)
出处
《西安工程大学学报》
CAS
2020年第5期87-92,共6页
Journal of Xi’an Polytechnic University
基金
柯桥纺织产业创新项目(19KQYB23)。
关键词
密度峰值
电商网络
社区检测
结点属性
直接邻居
间接邻居
density peaks
e-commerce network
community detection
node attribute
direct neighbors
indirect neighbors