Reply To: Help with hillclimb
Just a thought, or more perhaps a question… For the more advanced challenges, where white space has been removed, The normal gram (bigram, trigram etc.) frequencies, from plain English text, will be skewed by the ‘false’ letter groups generated from adjacent words. Do you produce your own counts, from text with white space removed?
My hill climbing method wasn’t very successful, but having previously used a Nelder–Mead downhill simplex method for minimum location in multidimensional space, we’d re-run the method with randomized starting conditions to weedle out local minima.