Skip to content

Clustering algorithm published on Science in 2014 is implemented in C++

License

Notifications You must be signed in to change notification settings

angusjch/cluster-science2014

 
 

Repository files navigation

=================== cluster-science2014

Clustering algorithm appeared on Science in 2014 is implemented in C++

/************************************************************

FileName: cluster_sci14.cpp

Author: Yunfei WANG

E-mail: wangyunfeiysm@163.com

Date: Nov.23,2014

***********************************************************/

OPTIONS

--input     Requests that all the samples are stored in the given file
            in which each row is a sample.
--clusters  Specify the number of clusters.
            If clusters<=0, the program will estimate the appropriate
            number of clusters automatically.
--withlabel Specify whether the last column in the input files are labels.
            0-the last column is also one of the features.
            1-the last column indicates the label of each sample.
--mode      Specify the approach to compute density.
            0-Gauss Kernel(default)
            1-Cutoff Kernel
            2-K Nearest Neighbors
--neighbors Specify the number of nearest neighbors, which is
            used to estiamte the density of each sample. It works
            only when the option '--mode 2'  is used.(default 5)
--metric    Specify the metric used to compute distance two samples.
            0-Euclidean(default)
            1-Cosine
--output    Specify the name of output files(default:the same as inputfile).
            The file named output.decisiongraph stores the decision graph,
            the two columns correspond to rho and delta respectively for each sample.
            The file named output.result stores the clustering result,
            in which the first column indicates the index of each sample
            and the second column indicate the index of its cluster.
--help

About

Clustering algorithm published on Science in 2014 is implemented in C++

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • C++ 81.2%
  • MATLAB 18.8%