基于矩阵加权关联规则的区间模糊C均值聚类

提出一种基于矩阵加权关联规则的区间模糊C均值聚类算法。根据支持度和可信度对矩阵构造关联规则，在关联规则的基础上进行区间模糊C均值聚类。由样本数量的大小来调整区间的影响因子a以达到最优聚类。该算法在解决小型文本时精度优于传统算法(如k-means)，在解决多维数据时效率较理想。理论和实验表明，该算法可以在一定程度上提高聚类结果的质量和算法效率。

Interborough Fuzzy C-Means Clustering Based on Matrix-weighted Association Rule

An algorithm of interborough fuzzy C-means clustering based on matrix-weighted association rule is presented. The method constructs the association rule about the matrix by supports and trusted degree, and conducts the interborough fuzzy C-means clustering based on association rule. It can adjust the affecting factors of interval data a, according to the size of swatch data, so that it can attain optimal cluster. It has much accuracy than the traditional algorithm with fewer text data(eg k-means), and efficiency is improved with multi dimensional data. Theoretic analysis and experimental demonstrations show that the algorithm outperforms existing algorithm in clustering quality and efficiency.