ch.akuhn.matrix.eigenvalues.Eigenvalues Maven / Gradle / Ivy
package ch.akuhn.matrix.eigenvalues;
import ch.akuhn.matrix.Matrix;
import ch.akuhn.matrix.Vector;
/**
* The eigen-decomposition of a matrix.
*
* @author Adrian Kuhn
*
*/
public class Eigenvalues {
/**
* The eigenvalues
*/
public double[] value;
/**
* The eigenvectors
*/
public Vector[] vector;
protected int n;
protected int nev;
/**
* Construct with the given dimensions
*
* @param n
*/
public Eigenvalues(int n) {
this.n = n;
this.nev = n;
}
/**
* Get an object that can compute the eigendecomposition of the given
* matrix. If the matrix has fewer than 10 columns, this will be an
* {@link AllEigenvalues}, otherwise it will be a {@link FewEigenvalues}.
*
* @param A
* @return the object to compute the eigen decomposition
*/
public static Eigenvalues of(Matrix A) {
if (A.columnCount() == 0) {
final Eigenvalues eigen = new Eigenvalues(0);
eigen.value = new double[0];
eigen.vector = new Vector[0];
return eigen;
}
if (A.columnCount() == 1) {
final Eigenvalues eigen = new Eigenvalues(0);
eigen.value = new double[] { A.get(0, 0) };
eigen.vector = new Vector[] { Vector.from(1.0) };
return eigen;
}
if (A.columnCount() < 10)
return new AllEigenvalues(A);
return FewEigenvalues.of(A);
}
/**
* Configure to compute the largest nev values/vectors.
*
* @param nev
* @return this
*/
public Eigenvalues largest(int nev) {
this.nev = nev;
return this;
}
/**
* Run the decomposition algorithm. Subclasses should override as necessary.
*
* @return this
*/
public Eigenvalues run() {
return this;
}
/**
* @return the total number of possible eigen vectors (i.e. the number of
* rows of the input)
*/
public int getN() {
return n;
}
}