Wang Xiaoyan, Bai Yanping
School of Information and Communication Engineering, North University of China, Taiyuan, 030051 People's Republic of China.
School of Science, North University of China, Taiyuan, 030051 People's Republic of China.
Springerplus. 2016 Sep 27;5(1):1665. doi: 10.1186/s40064-016-3329-4. eCollection 2016.
The global -means algorithm is an incremental approach to clustering that dynamically adds one cluster center at a time through a deterministic global search procedure from suitable initial positions, and employs -means to minimize the sum of the intra-cluster variances. However the global -means algorithm sometimes results singleton clusters and the initial positions sometimes are bad, after a bad initialization, poor local optimal can be easily obtained by -means algorithm. In this paper, we modified the global -means algorithm to eliminate the singleton clusters at first, and then we apply MinMax -means clustering error method to global -means algorithm to overcome the effect of bad initialization, proposed the global Minmax -means algorithm. The proposed clustering method is tested on some popular data sets and compared to the -means algorithm, the global -means algorithm and the MinMax -means algorithm. The experiment results show our proposed algorithm outperforms other algorithms mentioned in the paper.
全局均值算法是一种用于聚类的增量式方法，它通过从合适的初始位置进行确定性全局搜索过程，每次动态添加一个聚类中心，并采用均值法来最小化聚类内方差之和。然而，全局均值算法有时会产生单例聚类，并且初始位置有时不佳，在进行了糟糕的初始化之后，均值算法很容易得到较差的局部最优解。在本文中，我们首先修改了全局均值算法以消除单例聚类，然后将最小最大均值聚类误差方法应用于全局均值算法以克服糟糕初始化的影响，提出了全局最小最大均值算法。所提出的聚类方法在一些流行数据集上进行了测试，并与均值算法、全局均值算法和最小最大均值算法进行了比较。实验结果表明，我们提出的算法优于本文中提到的其他算法。
