quickstart.txt

Java方面的数值算法

Once you have downloaded the library, you can start using it in you Java 
programs. For now this guide features some basis sample programs that should
get you under way.

+----------------------------------------------+
| 1. Creating a dataset                        |
| 2. Creating and using a clustering algorithm |
+----------------------------------------------+

1. Creating a dataset
=====================
The code snippter below will create a two-dimensional dataset with four clusters of 
objects, one in each corner.

Dataset data = new SimpleDataset();
double small=1.0f/4.0f;
double large=3.0f/4.0f;
Random rg = new Random(System.currentTimeMillis());
for (int i = 0; i < itemsPerCluster; i++) {
    // lower left
	double[] vec1 = {  ((rg.nextGaussian() * clusterSpread)  + space * small),
    	 ((rg.nextGaussian() * clusterSpread)  + space * small) };
    data.addInstance(new SimpleInstance(vec1));
	// upper left
    double[] vec2 = {  ((rg.nextGaussian() * clusterSpread)  + space * small),
    	 ((rg.nextGaussian() * clusterSpread)  + space * large) };
    data.addInstance(new SimpleInstance(vec2));
	// lower righ
    double[] vec3 = {  ((rg.nextGaussian() * clusterSpread)  + space * large),
    	 ((rg.nextGaussian() * clusterSpread)  + space * small) };
    data.addInstance(new SimpleInstance(vec3));
    // upper right
    double[] vec4= {  ((rg.nextGaussian() * clusterSpread)  + space * large),
    	 ((rg.nextGaussian() * clusterSpread)  + space * large) };
    data.addInstance(new SimpleInstance(vec4));
}

2. Creating and using a clustering algorithm
============================================
The following code snippet will take as input the dataset created in the 
previous section, or any other dataset. 

XMeans km = new XMeans();
Dataset[] clusters = km.executeClustering(data);

In the array clusters, you find the different clusters that are the output of 
the algorithm.