protected override LogisticRegressionResult DoCompute(LogisticRegressionParameters parameters)
{
Func <LogisticRegressionParameters, Matrix, Matrix, Matrix> computeTheta;
if (parameters.Lambda == 0)
{
computeTheta = ComputeTheta;
}
else
{
computeTheta = ComputeThetaRegularized;
}
var theta = parameters.ThetaInit;
var costs = new List <double>();
var x = parameters.X;
var result = new LogisticRegressionResult();
x = MatrixUtils.Concatenate(MatrixUtils.Ones(parameters.M, 1), x);
costs.Add(ComputeCost(x, parameters.Y, theta, parameters.M));
for (int i = 0; i < parameters.IterationCount; i++)
{
theta = computeTheta(parameters, x, theta);
costs.Add(ComputeCost(x, parameters.Y, theta, parameters.M));
}
result.Costs = costs;
result.Theta = theta;
return(result);
}
protected override Matrix DoCompute(SolveEquationByGaussParameter parameters)
{
var ab = MatrixUtils.Concatenate(parameters.A, parameters.b);
var M = ab.ToMatrixDouble();
// pivoting
for (int col = 0; col + 1 < parameters.m; col++)
{
if (Math.Abs(M[col, col]) < parameters.EPSILON) // check for zero coefficients
{
// find non-zero coefficient
int swapRow = col + 1;
for (; swapRow < parameters.m; swapRow++)
{
if (Math.Abs(M[swapRow, col]) > parameters.EPSILON)
{
break;
}
}
if (Math.Abs(M[swapRow, col]) > parameters.EPSILON) // found a non-zero coefficient?
{
// yes, then swap it with the above
double[] tmp = new double[parameters.m + 1];
for (int i = 0; i < parameters.m + 1; i++)
{
tmp[i] = M[swapRow, i]; M[swapRow, i] = M[col, i]; M[col, i] = tmp[i];
}
}
else
{
throw new Exception("The matrix has no unique solution");
}
}
}
// elimination
for (int sourceRow = 0; sourceRow + 1 < parameters.m; sourceRow++)
{
for (int destRow = sourceRow + 1; destRow < parameters.m; destRow++)
{
double df = M[sourceRow, sourceRow];
double sf = M[destRow, sourceRow];
for (int i = 0; i < parameters.m + 1; i++)
{
M[destRow, i] = M[destRow, i] * df - M[sourceRow, i] * sf;
}
}
}
// back-insertion
for (int row = parameters.m - 1; row >= 0; row--)
{
double f = M[row, row];
if (Math.Abs(f) < parameters.EPSILON)
{
throw new Exception("");
}
for (int i = 0; i < parameters.m + 1; i++)
{
M[row, i] /= f;
}
for (int destRow = 0; destRow < row; destRow++)
{
M[destRow, parameters.m] -= M[destRow, row] * M[row, parameters.m];
M[destRow, row] = 0;
}
}
return(new Matrix(M).GetColumn(parameters.n - 1));
}
