I want to do multidimensional fits using singular value decomposition method.

The issue I need to handle are

I am reading a log file for my raw data which are in a vector of string data type. This data I want to use in the functions provided by Fitsvd() and funks(). Can anyone please tell me how to do it?
I am trying to do the facts mentioned above but I am encountering with many untraced errors leading me to system files.

Note:

1. I have minimum of 3 independent(x1, x2, x3) variables and one dependent variable(y) and 2 basis function.
2. I am a new C++ programmer and I am working with Visual C++ 2005.
3. In case if somebody has better idea than the one mentioned in the book, all ideas are invited.


Please help me!!

Merci
:)