There's a real speech as source text. The statistical distribution of the words is analyzed by a markov chain (char or word-based) which is then constantly producing new sentences; these seem to be grammatically correct but have of course some extremely boring or humoristic side-effects :)
this is absolutely fantastic. can we be friends?!
jamthehut 4 years ago
totally impressed!!!
tobsenteque 4 years ago
yes, indeed :D
There's a real speech as source text. The statistical distribution of the words is analyzed by a markov chain (char or word-based) which is then constantly producing new sentences; these seem to be grammatically correct but have of course some extremely boring or humoristic side-effects :)
audiocommander 4 years ago
great stuff.. am i correct to understand you implemented likely hood of the next word via markovchain?
ColdInferno 4 years ago