subcluster.h

Scalable k-means software and test datasets This package (a Unix tar file, gzipped) contains the sou

/* Scalable K-means clustering software
Copyright (C) 2000  Fredrik Farnstrom and James Lewis

This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; either version 2
of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.


See the file README.TXT for more information.

*/

/* subcluster.h */

class Singleton;

class Subcluster
{
	float *sum, *sumSquared;
	float *mean;
        float *stdDev;

	long numPoints;
	float sqrtNumPoints;

	int dimensions;

	char modified;

	Subcluster *next, *previous;
	Subcluster *clusterAssociation;
public:
	Subcluster(int dim);

	~Subcluster();

	static long getAllocated(void);

	void clear(void);

	// Update the sufficient statistics using point x.
	void addPoint(Singleton *x);

	// Update the sufficient statistics with cluster x.
	void addCluster(Subcluster *x);
	void subCluster(Subcluster *x);

	// Measure Mahalanobis distance from x to the mean of the subcluster.
	float Mahalanobis(float *x);
	float Mahalanobis(Singleton *p);
	float Mahalanobis(Subcluster *c);

	Subcluster *getPrevious(void)
	{	return previous;	}

	Subcluster *getNext(void)
	{	return next;	}

	void setPrevious(Subcluster *p)
	{	previous = p;	}

	void setNext(Subcluster *n)
	{	next = n;	}

	long getMemUsed(void)
	{	return sizeof(*this) + 4*dimensions*sizeof(float);
	}

	void associate(Subcluster *cluster)
	{	clusterAssociation = cluster;	}

	Subcluster *getAssociation(void)
	{	return clusterAssociation;	}

	float distanceSquared(Singleton *c);
	float distanceSquared(Subcluster *c);

	float updateMean(void);

        void computeStdDev(void);

	// Perturb the cluster mean within its confidence intervals, either
	// towards (if away == 0) or away from point p, and return distance
	// squared between new cluster mean and p. The cluster is left unchanged.
	float perturb(Singleton *p, int away, float stdDev);

	int isTight(float beta);

	long getNumPoints(void)
	{	return numPoints;	}

	void print(FILE *f = stdin);
	void printConfidence(FILE *f = stdin);
};

class MainCluster : public Subcluster
{
public:
	MainCluster(int dim) : Subcluster(dim)
	{	}

//	int numNewPoints;
	int id;
};

/* End of file subcluster.h */