public static Matrix2D Process(Matrix2D _mat, boolean removeNZRows) {
MatrixDescriptor desc = _mat.getMatrixDescriptor();
// Calculates the mean
System.out.println("Calculating Mean and Standar Deviation");
Matrix2D meanSDMat = RateOfMean.ColumnMeanAndStandardDeviation(_mat, removeNZRows);
// Calculate the standard score
Matrix2D standardScore = new Matrix2D(desc);
int numColumns = desc.getNumCols();
for (int column = 0; column < numColumns; column++) {
System.out.println("Processing column " + column + " of " + numColumns);
float _mean = meanSDMat.GetElement(RateOfMean.Mean, desc.getColumnAt(column));
float _sd = meanSDMat.GetElement(RateOfMean.StandardDeviation, desc.getColumnAt(column));
boolean denZero = Float.compare(0, _sd) == 0;
for (int row = 0; row < desc.getNumRows(); row++) {
float z = 0;
if (!denZero) z = (_mat.GetElement(row, column) - _mean) / _sd;
standardScore.SetElement(desc.getRowAt(row), desc.getColumnAt(column), z);
}
}
return standardScore;
}
public static void main(String[] arg) {
Matrix2D object_1 = new Matrix2D();
int[][] matrix_1 = {{1, 2}, {3, 4}};
int[][] matrix_2 = {{1, -5}, {3, 4}};
int[][] array = object_1.add(matrix_1, matrix_2);
for (int[] row : array) {
for (int colum : row) System.out.print(colum + "\t");
System.out.println();
}
int[][] multt = object_1.mult(matrix_1, matrix_2);
System.out.println("\nMult:");
for (int[] row : multt) {
for (int colum : row) System.out.print(colum + "\t");
System.out.println();
}
}
public static void Test3D() {
MatrixDescriptor testDesc = new MatrixDescriptor(InfoType.USER, InfoType.ARTIFACT);
testDesc.AddColDesc("circumf");
testDesc.AddColDesc("area_cy");
testDesc.AddColDesc("area_co");
testDesc.AddRowDesc("Alice");
testDesc.AddRowDesc("Carlos");
Matrix2D testMat = new Matrix2D(testDesc);
testMat.SetElement("Alice", "circumf", 2);
testMat.SetElement("Alice", "area_cy", 4);
testMat.SetElement("Alice", "area_co", 6);
testMat.SetElement("Carlos", "circumf", 6);
testMat.SetElement("Carlos", "area_cy", 8);
testMat.SetElement("Carlos", "area_co", 2);
Matrix2D layer2 = new Matrix2D(testDesc);
layer2.SetElement("Alice", "circumf", 7);
layer2.SetElement("Alice", "area_cy", 4);
layer2.SetElement("Alice", "area_co", 2);
layer2.SetElement("Carlos", "circumf", 2);
layer2.SetElement("Carlos", "area_cy", 9);
layer2.SetElement("Carlos", "area_co", 4);
Matrix3D _cube = new Matrix3D();
_cube.AddLayer(Database.CUBE_DEPTH_PREFIX + 0, testMat);
_cube.AddLayer(Database.CUBE_DEPTH_PREFIX + 1, layer2);
System.out.println();
System.out.println("z-score cube");
Matrix3D ssc = StandardScore.Process(_cube, Algorithm.CPU);
ssc.Debug();
}
private static Matrix2D SquaredVariance(Matrix2D _mat, Matrix2D _mean) {
MatrixDescriptor _meanDesc = _mean.getMatrixDescriptor();
MatrixDescriptor _matDesc = _mat.getMatrixDescriptor();
Matrix2D squaredVariance = new Matrix2D(_matDesc);
for (int column = 0; column < _matDesc.getNumCols(); column++) {
float columnMean = _mean.GetElement(0, column);
for (int row = 0; row < _matDesc.getNumRows(); row++) {
float res = _mat.GetElement(row, column) - columnMean;
squaredVariance.SetElement(_matDesc.getRowAt(row), _matDesc.getColumnAt(column), res * res);
}
}
return squaredVariance;
}
public static Matrix3D Process(Matrix3D _cube, Definitions.Algorithm _algorithm) {
Matrix3D res = new Matrix3D();
Set<Map.Entry<String, Matrix2D>> records = _cube.getIterator();
// Mean and standard deviation
System.out.println("Processing Mean and SD");
Matrix2D _meanMat = RateOfMean.ColumnDepthMeanAndStandardDeviation(_cube, true, _algorithm);
System.out.println("Processing layers...");
int c = 0;
boolean print = false;
boolean print2 = false;
for (Map.Entry<String, Matrix2D> entry : records) {
Matrix2D layer = entry.getValue();
MatrixDescriptor layerDesc = layer.getMatrixDescriptor();
Matrix2D layerSS = new Matrix2D(layerDesc);
// Matrix2D layerSS = layer.DoGPUStandardScore(_meanMat);
for (int col = 0; col < layerDesc.getNumCols(); col++) {
String colName = layerDesc.getColumnAt(col);
float _mean = _meanMat.GetElement(RateOfMean.Mean, layerDesc.getColumnAt(col));
float _sd = _meanMat.GetElement(RateOfMean.StandardDeviation, colName);
for (int row = 0; row < layerDesc.getNumRows(); row++) {
float _vv = layer.GetElement(row, col);
if (_vv >= 8 && print == false) {
print = true;
float _ss = (layer.GetElement(row, col) - _mean) / _sd;
System.out.println(_ss);
}
layerSS.SetElement(
layerDesc.getRowAt(row),
layerDesc.getColumnAt(col),
(layer.GetElement(row, col) - _mean) / _sd);
}
}
if (print == true && print2 == false) {
print2 = true;
layerSS.Debug();
}
res.AddLayer(entry.getKey(), layerSS);
System.out.println("Standard Score of matrix " + c + " of " + records.size());
c++;
}
return res;
}
public static void Test2D() {
MatrixDescriptor testDesc = new MatrixDescriptor(InfoType.USER, InfoType.ARTIFACT);
testDesc.AddColDesc("circumf");
testDesc.AddColDesc("area_cy");
testDesc.AddColDesc("area_co");
testDesc.AddColDesc("draw");
testDesc.AddRowDesc("Alice");
testDesc.AddRowDesc("Carlos");
testDesc.AddRowDesc("Bob");
Matrix2D testMat = new Matrix2D(testDesc);
testMat.SetElement("Alice", "circumf", 1);
testMat.SetElement("Alice", "area_cy", 1);
testMat.SetElement("Alice", "area_co", 1);
testMat.SetElement("Alice", "draw", 14);
testMat.SetElement("Carlos", "circumf", 5);
testMat.SetElement("Carlos", "area_cy", 3);
testMat.SetElement("Carlos", "area_co", 3);
testMat.SetElement("Carlos", "draw", 5);
testMat.SetElement("Bob", "circumf", 10);
testMat.SetElement("Bob", "area_cy", 1);
testMat.SetElement("Bob", "area_co", 1);
testMat.SetElement("Bob", "draw", 1);
System.out.println("DM Matrix");
testMat.Debug();
System.out.println();
System.out.println("z-score Matrix");
Matrix2D ss = StandardScore.Process(testMat, true);
ss.Debug();
}
