基于大型数据集的聚类算法研究

The Research on Clustering Algorithm for Large Datasets

就精确系数不算太严格的情况而言,针对各种大型数据集,通过对比各种聚类算法,提出了一种部分优先聚类算法。然后在此基础之上分析研究聚类成员的产生过程与聚类融合方式,通过设计共识函数并利用加权方式确定类中心,在部分优先聚类算法的基础上进行聚类融合,从而使算法的计算准度加以提升。通过不断的实验,我们可以感受到优化之后算法的显著优势,这不仅体现在其可靠性,同时在其稳定性以及扩展性、鲁棒性等方面都得到了很好的展现。

For all kinds of large datasets,a kind of part priority clustering algorithm is proposed based on the comparison of various clustering algorithms.Then on the basis of the analysis and research on the generation process of cluster members and the ensemble method,by designing the consensus function and using the weighted method to determine the class center,based on the partial priority clustering algorithm for clustering ensemble,so as to improve the accuracy of the algorithm.Through continuous experiments,we can feel the significant advantage of the optimiza-tion algorithm,which is not only reflected in its reliability,but also in its stability and scalability,robustness and so on have been a good show.