摘要： 介绍了模糊C 均值聚类算法的实现途径，并针对这种算法对初值敏感的缺点，研究实现了一种已被应用的模糊C 均值的自适应算法，即将用于聚类分析有效性评价的混合F 统计量与算法结合在一起的一种改进算法。该算法能够自动确定最佳聚类数目，避免在聚类数目的选取上存在的主观性，解决了聚类中的全局最优问题，提高了算法的可靠程度。模糊聚类分析方法及有效性评估已在某市应急指挥系统项目中得到了应用。

关键词: 模糊聚类； 模糊 C 均值；应急事件响应

Abstract: This paper is about the realization of fuzzy C-means cluster algorithm(FCM). It realizes an self-adapted FCM, which is put into practice, considering the algorithm's shortcoming of initial value sensitive. That is, it is an improved algorithm combining FCM together with mixed F-statistic which is used to evaluate the effectiveness of cluster analysis. It makes it automatic to determine the best number of classes to be divided, avoiding subjective affection on this determination. It solves the problem of being perfect as a whole, and increases its reliability. This algorithm is implemented in a city’s emergency events response application

Key words: Fuzzy cluster; Fuzzy C-means; Emergency events response