一种基于PSO-FCM的网络虚拟环境信息推荐算法

An algorithm of information recommendation based on PSO-FCM in network virtual environments

提出一种新的基于PSO-FCM用户聚类的信息推荐算法.PSO-FCM算法结合了PSO与FCM的优点,避免了FCM算法对初始值、噪声数据敏感与PSO容易陷入局部最优等缺点.为增强聚类效果,在PSO中设计了一个基于双目标的粒子适应度评价函数,最后用标准数据集与模拟数据集对推荐算法进行实验测试.结果表明,所提的基于PSO-FCM的信息推荐算法有良好的表现.

A new information recommendation algorithm based on PSO-FCM users clustering would be proposed in this paper.The PSO-FCM,i.e.Fuzzy C-Means（FCM） clustering based on particle swarm optimization（PSO）,combines the respective advantages of PSO and FCM,which could prevent from those defects,for FCM to be susceptible to initial value and noisy data,and PSO to be easily falling into local optimum and so on.In order to improve clustering effect,we design the particle fitness function based on dual-objectives（intra-class distance and inter-class distance） in PSO.Finally,the standard data set and simulation data set are applied to test the personalized information recommendation algorithm based on PSO-FCM users clustering.The experimental result shows that this algorithm was of good performances.