Rprop implementation in Octave/MATLAB for efficient gradient-based optimization.
- M. Riedmiller, H. Braun (1992) Rprop: A Fast Adaptive Learning Algorithm. In International Symposium on Computer and Information Science VII. pp. 279 - 286. Antalya, Turkey.
Add the source code to your Matlab path. You can get help by typing
help rprop
or run the demo to find the minimum of a polynomial:
testrprop
The objective must be given as a Matlab function of the following form:
[y, g] = func(x, P1, P2, ...)
where y is the function value and g is the gradient wrt the parameters x. x are the parameters to be optimized and P1, P2, ... are additional parameters (optional).