// Return the value for vector c which best fits the linear model y = Xc.
// X is a matrix given by its row vectors: each row corresponds to a y value;
// columns within the row correspond to coefficients for parameters in c.
func Linear(y vec.Vector, X []vec.Vector) vec.Vector {
n := C.size_t(len(y)) // number of observations
p := C.size_t(len(X[0])) // number of parameters
c, gslY := C.gsl_vector_alloc(p), C.gsl_vector_alloc(n)
cov, gslX := C.gsl_matrix_alloc(p, p), C.gsl_matrix_alloc(n, p)
vecToGSL(y, gslY)
matrixToGSL(X, gslX)
chisq := C.double(0)
work := C.gsl_multifit_linear_alloc(n, p)
C.gsl_multifit_linear(gslX, gslY, c, cov, &chisq, work)
C.gsl_multifit_linear_free(work)
result := vecFromGSL(c)
C.gsl_matrix_free(gslX)
C.gsl_matrix_free(cov)
C.gsl_vector_free(gslY)
C.gsl_vector_free(c)
return result
}
func (x *GslMultifitLinearWorkspace) Dispose() {
C.gsl_multifit_linear_free((*C.gsl_multifit_linear_workspace)(unsafe.Pointer(x.Ptr())))
}