一种改进的多视图聚类集成算法

Improved Multi-view Clustering Ensemble Algorithm

近年来,针对大数据的数据挖掘技术和机器学习算法研究变得日趋重要。在聚类领域,随着多视图数据的大量出现,多视图聚类已经成为了一类重要的聚类方法。然而,大多数现有的多视图聚类算法受算法参数设置、数据样本等影响,具有聚类结果不稳定、参数需要反复调节等缺点。基于多视图K-means算法和聚类集成技术,提出了一种改进的多视图聚类集成算法,其提高了聚类的准确性、鲁棒性和稳定性。其次,由于单机环境下的多视图聚类算法难以对海量的数据进行处理,结合分布式处理技术,实现了一种分布式的多视图并行聚类算法。实验证明,并行算法在处理大数据时的时间效率有很大提升,适合于大数据环境下的多视图聚类分析。

In recent years, data mining and machine learning algorithms for big data become increasingly important. In the clustering,with the appearance of multi-view data, multi-view clustering has become an important clustering method. However,many existing multi-vlew clustering algorithms are easily affected by parameter setting and dataset itself, so the clustering results are usually unstable. To overcome this problem, we presented a new multi-view clustering ensem- ble algorithm based on the multi-view K-means clustering algorithm in this paper. This algorithm uses ensemble tech- nique to improve the multi-view K-means algorithm performance, increasing the accuracy, robustness, and stability of clustering results. It is well known that one single computer cannot process too much data, because one computer has the limited computation resources. To improve the efficiency of multi-view clustering, we implemented a distributed multi-view clustering ensemble algorithm based on distributed processing technology. Experimental results show that the proposed approach has higher efficiency when processing large dataset, and it is suitable for multi-view clustering in big data environment.