commons-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Shubham Jindal (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (MATH-1435) Implement cKMeans as a clustering algorithm
Date Thu, 09 Nov 2017 16:03:00 GMT

     [ https://issues.apache.org/jira/browse/MATH-1435?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Shubham Jindal updated MATH-1435:
---------------------------------
    Description: 
cKMeans implementation has been described here
https://cran.r-project.org/web/packages/Ckmeans.1d.dp/index.html and https://journal.r-project.org/archive/2011-2/RJournal_2011-2_Wang+Song.pdf

The algorithm described here is O(kn^2) where k: number of clusters and n: number of 1D points.
But, there exists an efficient implementation in later versions of cKMeans which is O(knlogn)

cKMeans is faster than kMeans and also deterministic in nature. It is supposed to be one of
the best clustering algorithms for clustering 1D points

  was:
cKMeans implementation has been described here
https://cran.r-project.org/web/packages/Ckmeans.1d.dp/index.html and https://journal.r-project.org/archive/2011-2/RJournal_2011-2_Wang+Song.pdf

The algorithm described here is O(kn^2) where k: number of clusters and n: number of 1D points.
But, there exists an efficient implementation in later versions of cKMeans which is O(knlogn)

cKMeans is faster than kMeans and also deterministic in nature. cKMeans is supposed to be
one of the best clustering algorithms for clustering 1D points


> Implement cKMeans as a clustering algorithm
> -------------------------------------------
>
>                 Key: MATH-1435
>                 URL: https://issues.apache.org/jira/browse/MATH-1435
>             Project: Commons Math
>          Issue Type: New Feature
>            Reporter: Shubham Jindal
>
> cKMeans implementation has been described here
> https://cran.r-project.org/web/packages/Ckmeans.1d.dp/index.html and https://journal.r-project.org/archive/2011-2/RJournal_2011-2_Wang+Song.pdf
> The algorithm described here is O(kn^2) where k: number of clusters and n: number of
1D points. But, there exists an efficient implementation in later versions of cKMeans which
is O(knlogn)
> cKMeans is faster than kMeans and also deterministic in nature. It is supposed to be
one of the best clustering algorithms for clustering 1D points



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message