基于最小生成树的不平衡数据集聚类算法

Clustering Algorithm for Unbalanced Datasets Based on Minimum Spanning Tree

采用传统不平衡数据集聚类算法直接对数据集编码树进行构建,而未对数据集密度特征进行提取,造成传统算法数据聚类效果差,因此提出了基于最小生成树的不平衡数据集聚类算法。先利用数据区域密度的敏感性,提取数据密度特征,再利用提取的数据集密度特征构建编码树,并计算不平衡聚类状态下的数据集,最后基于最小生成树实现不平衡数据集聚类。设计对比实验,实验结果表明该研究算法聚类效果最好,具有研究价值。

The traditional imbalanced dataset clustering algorithm is used to directly construct the dataset encoding tree without extracting the density features of the dataset,resulting in poor clustering performance of the traditional algorithm.Therefore,a minimum spanning tree based imbalanced dataset clustering algorithm is proposed.Firstly,the sensitivity of data region density is utilized to extract data density features.Then,an encoding tree is constructed using the extracted density features of the dataset,and the dataset in imbalanced clustering state is calculated.Finally,imbalanced dataset clustering is achieved based on the minimum spanning tree.Design comparative experiments,and the experimental results show that the clustering effect of this research algorithm is the best,which has research value.