Logarithmic Growth in Biological Processes

The aim of our paper is to approximate the parameters of a generic logarithmic function that can be used to model various biological processes. We are considering the Gauss-Newton algorithm for solving the non-linear least squares problem and we are proposing a method for selecting the initial choice for the algorithm. This method proves to considerably increase the convergence probability of the Gauss-Newton algorithm applied for our function compared to its convergence probability when choosing a purely heuristic initial approximation.