基于余弦相似性的自适应权重的改进FCM算法  

Improved FCM algorithm with adaptive weights based on cosine similarity

作　　者：胡建华[1] 尹慧琳 HU Jianhua;YIN Huilin(College of Science,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学理学院,上海200093

出　　处：《智能计算机与应用》2021年第7期73-79,共7页Intelligent Computer and Applications

基　　金：国家自然科学基金(61873169)。

摘　　要：模糊C-均值聚类算法(FCM)是一种经典的聚类算法,主要通过迭代更新隶属度和聚类中心来提高聚类的有效性。FCM算法的性能主要通过类内紧性和类间分离性来评价,但其既依赖于初始聚类中心,也对噪声非常敏感。考虑到每个数据点和每个聚类中心对目标函数的不同重要性,本文提出了一种具有自适应权重的改进FCM聚类算法(Hybrid FCM)。主要贡献:将2个具有自适应指数p和q的自适应权向量ψ和φ引入FCM的目标函数,以体现不同数据点和聚类中心的重要性;为提高聚类性能,自适应指数p、q和模糊因子m采用粒子群优化算法(PSO)优化,新提出的聚类评价指标AWCVI作为PSO算法的适应度函数;迭代过程中利用余弦相似性对隶属度函数进行修正,提高算法的鲁棒性。实验表明,本文提出的算法能够有效地提高聚类效果。As a classical clustering algorithm,fuzzy C-means clustering algorithm(FCM)improves the effectiveness of clustering by updating membership and clustering centers.The performance of FCM algorithm is mainly evaluated by intra cluster compactness and inter cluster separation.But FCM algorithm relies on the clustering centers and is very sensitive to noise.Considering the different importance of each data point and each cluster center,two adaptive weight vectorsψandφwith adaptive exponents p and q are introduced into the objective function of FCM,and an improved FCM clustering algorithm with adaptive weights is proposed;At the same time,the new clustering evaluation index AWCVI is used to optimize the parameters p,q and the fuzzy factor m,which is determined by the particle swarm optimization algorithm(PSO);The cosine similarity is used to modify the membership function in the iterative process to improve the robustness of the algorithm.Experimental results show that the proposed algorithm can effectively improve the clustering effect.

关 键 词：模糊C均值算法 自适应权重 余弦相似度 粒子群算法 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]