//! multi dimensional linear regression
public LinearRegression(List <List <double> > x, List <double> y)
{
reg_ = new LinearLeastSquaresRegression <List <double> >(x, y, linearFcts(x[0].Count));
}
public void calibrate()
{
int n = paths_.Count;
Vector prices = new Vector(n), exercise = new Vector(n);
int len = EarlyExerciseTraits <PathType> .pathLength(paths_[0]);
for (int i = 0; i < paths_.Count; i++)
{
prices[i] = pathPricer_.value(paths_[i], len - 1);
}
for (int i = len - 2; i > 0; --i)
{
List <double> y = new List <double>();
List <double> x = new List <double>();
//roll back step
for (int j = 0; j < n; ++j)
{
exercise[j] = pathPricer_.value(paths_[j], i);
if (exercise[j] > 0.0)
{
x.Add(pathPricer_.state(paths_[j], i));
y.Add(dF_[i] * prices[j]);
}
}
if (v_.Count <= x.Count)
{
coeff_[i] = new LinearLeastSquaresRegression <double>(x, y, v_).coefficients();
}
else
{
// if number of itm paths is smaller then the number of
// calibration functions -> no early exercise
coeff_[i] = new Vector(v_.Count);
}
for (int j = 0, k = 0; j < n; ++j)
{
prices[j] *= dF_[i];
if (exercise[j] > 0.0)
{
double continuationValue = 0.0;
for (int l = 0; l < v_.Count; ++l)
{
continuationValue += coeff_[i][l] * v_[l](x[k]);
}
if (continuationValue < exercise[j])
{
prices[j] = exercise[j];
}
++k;
}
}
}
// remove calibration paths
paths_.Clear();
// entering the calculation phase
calibrationPhase_ = false;
}
//! one dimensional linear regression
public LinearRegression(List <double> x, List <double> y)
{
reg_ = new LinearLeastSquaresRegression <List <double> >(argumentWrapper(x), y, linearFcts(1));
}
