【摘要】 分析了现有半监督FCM算法目标函数的物理意义和平衡系数α的选取,说明Stutz对Pedrycz目标函数的修改使半监督的物理意义更清楚,它在α=1,0时均退化为标准FCM算法,给出了修改后SS-FCM算法的交替求解过程。实验结果:(1)修改算法与Pedrycz算法有相同的半监督作用和清楚的物理解释;(2)对labeled样本采用FCM算法赋值比用随机数的收敛稳定性高;(3)优选的少量labeled样本,使用模糊协方差的SS-CFCM算法提高了聚类准确性和收敛速度。更多还原

【Abstract】 Analyze the physical interpretation of objective function of semi-supervised FCM algorithm and the coefficient α.Illustrate that Stutz’s modification to the objective function provided by Pedrycz is more clear,and when α=1,0,the SS-FCM degrades to FCM.Provide the corresponding alternatively optimizing algorithm of SS-FCM.The experimental results show that:(1) Modified algorithm has same semi-supervised function and has more clear physical interpretation.(2)Using FCM algorithm to as-sign membership for labeled samples is better than using random number(.3)SS-FCM with fuzzy covariance and a small number of good-selected labeled samples can effectively improve the accuracy and convergence rate.更多还原