get eigenvalue pca with java
Posted
by
Muhamad Burhanudin
on Stack Overflow
See other posts from Stack Overflow
or by Muhamad Burhanudin
Published on 2013-06-30T00:54:52Z
Indexed on
2013/10/31
3:54 UTC
Read the original article
Hit count: 311
I try use PCA to reduce dimention, and i use jama for help me using matrix. but, i got problem when get eigenvalue with jama. for example i hava 2 image dimention 100x100, then i create single matrix 2 image x (100x100). there is 20.000 pixel. and how to get reduction with eigenvalue? this is sample my code :
public static void main(String[] args) {
BufferedImage bi;
int[] rgb;
int R, G, B;
// int[] jum;
double[][] gray = new double[500][500] ;
String[] baris = new String[1000];
try {
//bi = ImageIO.read(new File("D:\\c.jpg"));
int[][] pixelData = new int[bi.getHeight() * bi.getWidth()][3];
int counter = 0;
for (int i = 0; i < bi.getHeight(); i++) {
for (int j = 0; j < bi.getWidth(); j++) {
gray[i][j] = getPixelData(bi, i, j);
// R = getR(bi, i, j);
//G = getG(bi, i, j);
//B = getB(bi, i, j);
//jum = R + G + B;
// gray[counter] = Double.toString(R + G + B / 3);
// System.out.println("Gray " +gray);
//for (int k = 0; k < rgb.length; k++) {
// pixelData[counter][k] = rgb[k];
// }
counter++;
}
}
} catch (IOException e) {
e.printStackTrace();
}
Matrix matrix = new Matrix(gray);
PCA pca = new PCA(matrix);
pca.getEigenvalue(6);
String n = pca.toString();
System.err.println("nilai n "+n);
//double dete = pcadete(matrix,3600);
}
private static int getPixelData(BufferedImage bi, int x, int y) {
int argb = bi.getRGB(y, x);
int r, g, b;
int gray;
int rgb[] = new int[]{
(argb >> 16) & 0xff, //red
(argb >> 8) & 0xff, //green
(argb) & 0xff //blue
};
r = rgb[0];
g = rgb[1];
b = rgb[2];
gray = (r + g + b) / 3;
System.out.println("gray: " + gray);
return gray;
}
when i show eigenvalue in this code :
PCA pca = new PCA(matrix);
pca.getEigenvalue(6);
String n = pca.toString();
System.err.println("nilai n "+n);
Result is :
nilai n PCA@c3e9e9
Can, u tell me what way to get eigenvalue and reduction dimension.
© Stack Overflow or respective owner