This program calculates the fitting parameters for a linear least-squares problem.
However, it is not a "simple" regression in a way of y over x, which is fitted by y_i = a + b * x_i,
but a fit of many y_ij over their resp. x_i, y_ij = a + b * c_j * x_i.
In other words, if you have values to be fitted that can be described by the same fitting parameters A and B and the slope only differs by a parameter, you can fit all together.
