public CPointDistributionModel principalComponentAnalysis(Size clippingSize, out long time)
{
long start = CSettings.currentTimeMillis();
// Get Covarian Matrix
CMatrix s = new CMatrix(_numPoints * 2, false);
for (int i = 0; i < _numSamples; i++)
{
CVector dx = new CVector(_trainingData[i].substractCopy(_mean));
CMatrix dxdxT = dx.multiplyMatrix(dx);
s.add(dxdxT);
}
s.divide(_numSamples);
// Calculate Eigen
CPointDistributionModel ret = new CPointDistributionModel(clippingSize, _mean, CEigen.calculate(s), 7, _connectivity);
long end = CSettings.currentTimeMillis();
time = end - start;
return(ret);
}
static void Main(string[] args)
{
CMatrix4x4 temp = new CMatrix4x4(
new CVector4(10, 5, 3, 7),
new CVector4(-2, 7, -10, 1),
new CVector4(0, 3, 1, 6),
new CVector4(2, -3, 0, -1));
CMatrix4x4 inv = temp.inverse();
CMatrix buff = new CMatrix(3, false);
buff[0][0] = 8.0 / 9; buff[0][1] = 1.0 / 2; buff[0][2] = 1.0 / 3;
buff[1][0] = 1.0 / 2; buff[1][1] = 1.0 / 3; buff[1][2] = 1.0 / 4;
buff[2][0] = 1.0 / 3; buff[2][1] = 1.0 / 4; buff[2][2] = 1.0 / 5;
CEigen res = CEigen.calculate(buff);
System.Console.WriteLine("T");
for (int i = 0; i < 3; i++)
{
for (int j = 0; j < 3; j++)
{
System.Console.Write(res.vectors[i, j] + " ");
}
System.Console.WriteLine();
}
System.Console.WriteLine("d");
for (int i = 0; i < 3; i++)
{
System.Console.WriteLine(res.values[i]);
}
CVector dx = new CVector(10);
dx[0] = 3.369;
dx[1] = 9.999;
dx[2] = 12;
dx[3] = 0.00458;
dx[4] = 398;
dx[5] = 0.0782;
dx[6] = 1.596;
dx[7] = 4.2118;
dx[8] = 6.2258;
dx[9] = 0.0795;
CMatrix ret = dx.multiplyMatrix(dx);
System.Console.WriteLine("dxdxT");
for (int i = 0; i < 10; i++)
{
for (int j = 0; j < 10; j++)
{
System.Console.Write(ret[i, j] + " ");
}
System.Console.WriteLine();
}
System.Console.In.ReadToEnd();
}