The parameter dimension is not relevant for the peak filtering, since you can apply a lowpass filter in any dimension. But the training method is. It is only implemented for Ascent and brute force, not for Genetic. For implementing your own algorithm, look in the manual for the parameters() function.