Package ffx.potential.utils

Class Clustering

java.lang.Object

ffx.potential.utils.Clustering

public class Clustering
extends Object

Cluster contains methods utilized in the Cluster.groovy file.

Author:

Aaron J. Nessler, Michael J. Schnieders

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static class	Clustering.Conformation	Class for cluster objects.

Constructor Summary

Constructors

Constructor	Description
Clustering()

Method Summary

Modifier and Type	Method	Description
static void	analyzeClusters(List<org.apache.commons.math3.ml.clustering.CentroidCluster<Clustering.Conformation>> clusters, List<Integer> repStructs, boolean verbose)	Analyze a list of CentroidClusters.
static List<org.apache.commons.math3.ml.clustering.CentroidCluster<Clustering.Conformation>>	hierarchicalClustering(List<double[]> distanceMatrix, double threshold)	This method performs hierarchical clustering on a distance matrix.
static List<org.apache.commons.math3.ml.clustering.CentroidCluster<Clustering.Conformation>>	iterativeClustering(List<double[]> distMatrix, int trials, double tolerance)	Perform an iterative clustering for a specified number of clusters.
static List<org.apache.commons.math3.ml.clustering.CentroidCluster<Clustering.Conformation>>	kMeansClustering(List<double[]> distMatrix, int maxClusters, int numTrials, long seed)	Perform a k-means clustering for a specified number of clusters.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

Clustering

public Clustering()

Method Details

kMeansClustering

public static List<org.apache.commons.math3.ml.clustering.CentroidCluster<Clustering.Conformation>> kMeansClustering(List<double[]> distMatrix,
 int maxClusters,
 int numTrials,
 long seed)

Perform a k-means clustering for a specified number of clusters.

Parameters:

distMatrix - Coordinate input serves as the data points.

maxClusters - Number of clusters to use (k).

numTrials - Number of trials for the Multi K-Means++ algorithm.

seed - The seed to use for clustering (-1 uses the current system time).

Returns:

The clusters.

hierarchicalClustering

public static List<org.apache.commons.math3.ml.clustering.CentroidCluster<Clustering.Conformation>> hierarchicalClustering(List<double[]> distanceMatrix,
 double threshold)

This method performs hierarchical clustering on a distance matrix. If the system isn't headless, a dendrogram is printed of the clustered results. A PDB file for the centroid of each cluster is saved.

Parameters:

distanceMatrix - A List of double[] entries that holds the distance matrix.

threshold - the distance used to separate clusters.

Returns:

Return a list of CentroidClusters.

iterativeClustering

public static List<org.apache.commons.math3.ml.clustering.CentroidCluster<Clustering.Conformation>> iterativeClustering(List<double[]> distMatrix,
 int trials,
 double tolerance)

Perform an iterative clustering for a specified number of clusters. Designed and tested by researchers at Takeda (see method authors).

Parameters:

distMatrix - Coordinate input serves as the data points.

trials - Number of iterations to perform clustering.

tolerance - RMSD cutoff to divide same values from different.

Returns:

The clusters. Created by:

analyzeClusters

public static void analyzeClusters(List<org.apache.commons.math3.ml.clustering.CentroidCluster<Clustering.Conformation>> clusters,
 List<Integer> repStructs,
 boolean verbose)

Analyze a list of CentroidClusters.

Parameters:

clusters - The List of CentroidClusters to analyze.

repStructs - Store a representative conformation for each cluster.

verbose - If true, use verbose printing.