Abstract: A self-adapting fuzzy c means clustering algorithm is proposed for clustering problems with interval data. On the one hand, based on the mid-point and half-width of interval value, the impact factor of interval width is introduced to control the influence on the clustering results by the length of the interval data. On the other hand, the impact on the clustering results by the data structure of the interval data can also be reduced by introducing a self-adapting coefficient. The validity of the algorithm can be demonstrated by synthetic data and Fish real data, and the clustering results are compared and analyzed.

Key words: interval data, fuzzy c means clustering, self-adapting coefficient, self-adapting fuzzy c means clustering

摘要： 针对区间型数据的聚类问题，提出一种自适应模糊c均值聚类算法。该算法一方面基于区间数的中点和半宽度，通过引入区间宽度的影响因子以控制区间大小对聚类结果的影响；另一方面通过引入一个自适应系数，以减少区间型数据的数据结构对聚类效果的影响。通过仿真数据和Fish真实数据验证了该算法的有效性，并对聚类结果进行比较和分析。

关键词: 区间型数据, 模糊c均值聚类, 自适应系数, 自适应模糊c均值聚类