新型模糊半监督加权聚类算法中的权值v

Feature weight vin novel semi-supervised fuzzy clustering algorithm with feature discrimination

为体现实际应用中不同特征对聚类性能的影响,提出一种模糊半监督加权聚类算法。采用聚类和特征加权同步进行的思想,将特征加权融合到聚类过程中,对不同约束条件下的权值对聚类准确率的影响进行分析,实时观测权值在聚类过程中的变化轨迹的目标。在UCI和自测数据集上的相关实验结果表明,权值能够有效提高算法的性能。

To reflect the weight effectiveness of various features on the performance of clustering,a semi-supervised fuzzy cluste-ring algorithm with feature discrimination （SFFD）was proposed,in which feature weighting was incorporated into the process of clustering and the clustering and feature weighting were performed synchronously.The influence of the weights on the accuracy of clustering was analyzed under various amounts of constraints.During the clustering process,the traj ectories of weights were observed in real time.Experimental results on UCI and self-testing datasets demonstrate that the weight can effectively improve the performance of the algorithm.