Louvain 算法与 K 均值聚类算法的比较研究

Comparative study of Louvain algorithm and K-means clustering algorithm

  • 摘要: 复杂网络是近年来新兴的研究领域,社区发现是其应用方向之一。对于现实数据集进行聚类 分析是数据挖掘的一个重要方法,但存在聚类分析效果不佳的情形。此时若引入相关性度量, 将数据集构建成复杂网络,便可使用社区发现方法对其进行处理。现有文献大多针对算法进 行改进,对两种方法的划分结果进行比较的研究较少。本文选取了社团划分中的 Louvain 算 法与聚类算法中的 K 均值聚类算法,首先对两种算法的理论进行比较,接着利用心脏病、肾 病患者数据构造复杂网络,比较了 Louvain 算法的社区划分结果与 K 均值聚类算法的聚类结 果,在正确划分率的评价标准下,Louvain 算法的社团划分结果优于 K 均值聚类算法的聚类 结果。

     

    Abstract: Complex network is a new research field in recent years, and community discovery is one of its application directions. Clustering analysis is an important method of data mining for real data sets, but the effect of clustering analysis is not good. At this time, if we introduce the correlation measure and construct the data set into a complex network, we can use the community discovery method to process it. Most of the existing literatures focus on the improvement of the algorithm, and there is less research on the comparison of the results of the two methods. In this paper, we select the Louvain algorithm in community partition and K-means clustering algorithm in clustering algorithm. Firstly, we compare the theories of the two algorithms, and then use the data of patients with heart disease and kidney disease to construct a complex network. We compare the community partition results of the Louvain algorithm and the clustering results of the K-means clustering algorithm, The result of community partition of Louvain algorithm is better than that of K-means clustering algorithm.

     

/

返回文章
返回