基于局部引力模型的聚类算法研究综述

A Review of Local Gravitation Clustering

局部引力模型(Local Gravitation Model,LGM)是2018年提出的用于数据聚类的方法,其区别于传统引力聚类模型之处在于计算局部中心量度的过程中同时使用了局部引力合力的模长和方向信息。近年来,研究人员针对局部引力模型进行了改进泛化,提出了一系列新型聚类算法。笔者就局部引力聚类算法的发展趋势进行综述,对相关研究方向作展望。

The Local Gravitation Model(LGM)in machine learning is proposed for data clustering in the year 2018.Unlike the traditional gravitational clustering model,the LGM takes both the magnitude and the direction of the local result force into consideration.Recently,some researchers published several papers that improve the performance of the LGM,with several new clustering algorithms proposed.This paper reviews the recent development of the LGM,and discusses the foreground of the LGM.