Skip to content

Gathering some unusual clustering demand , also include some optimization of popular algorithms

Notifications You must be signed in to change notification settings

laura-zhang-cn/-Clustering-Algorithm

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 

Repository files navigation

-Clustering-Algorithm

Gathering some unusual clustering demand , also include some optimization of popular algorithms 主要是处理一些特殊的聚类需求,或者 改进一些流行的聚类算法

1. 经纬度聚类 longitude-latitude-clustering
对经纬度聚类,目标是 固定半径内的点聚类,采用meanshift的逻辑+使用haversine距离算子。
若不固定聚类结果的半径,是基于密度可达的密集区域,则直接使用DBSCAN,metric参数设置为haversine即可 。

注意:
DBSCAN是不返回类中心的 ,且sklearn自带的 haversine距离算子要求传入的 [x,y]的顺序是[lat,lng]
meanshift原本是应用在图像识别领域,有固定的核函数,遂这里应用到经纬度时,需要自己构建代码,在迭代过程中加入haversine距离算子。

About

Gathering some unusual clustering demand , also include some optimization of popular algorithms

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages