private double[] gradient(double[] w)
{
double[] y = Outputs;
double p = Epsilon;
sizeI = 0;
for (int i = 0; i < y.Length; i++)
{
double d = z[i] - y[i];
// generate index set I
if (d < -p)
{
z[sizeI] = C[i] * (d + p);
I[sizeI] = i;
sizeI++;
}
else if (d > p)
{
z[sizeI] = C[i] * (d - p);
I[sizeI] = i;
sizeI++;
}
}
LinearNewtonMethod.subXTv(Inputs, biasIndex, I, sizeI, z, g);
for (int i = 0; i < w.Length; i++)
{
g[i] = w[i] + 2 * g[i];
}
return(g);
}
private double objective(double[] w)
{
double[] y = Outputs;
double p = Epsilon;
LinearNewtonMethod.Xv(Inputs, biasIndex, w, z);
double f = 0;
for (int i = 0; i < w.Length; i++)
{
f += w[i] * w[i];
}
f /= 2;
for (int i = 0; i < y.Length; i++)
{
double d = z[i] - y[i];
if (d < -p)
{
f += C[i] * (d + p) * (d + p);
}
else if (d > p)
{
f += C[i] * (d - p) * (d - p);
}
}
return(f);
}
private double[] hessian(double[] s)
{
LinearNewtonMethod.subXv(Inputs, biasIndex, I, sizeI, s, wa);
for (int i = 0; i < sizeI; i++)
{
wa[i] = C[I[i]] * wa[i];
}
LinearNewtonMethod.subXTv(Inputs, biasIndex, I, sizeI, wa, h);
for (int i = 0; i < s.Length; i++)
{
h[i] = s[i] + 2 * h[i];
}
return(h);
}