Use markov chains and nGrams to generate presidential speeches, based on previous ones.
A Markov chain is a process that determines a sequence of events (in our case, characters) based solely on the probability of an event given a few events prior to it. That probability is decribed in an nGram.
This algorithm takes in presidential speeches from the past, creates nGram representations for it, then generates text from those nGram models.
Run python tests.py
to visualize our quick demo.
from markov import *
model1 = build_model(tokensOrString, n) # n is the number of elements to base the nGram on.
model2 = build_model(tokensOrString2, n)
model = merge_models([model1, model2])
# seed is the initial token, where to start the generation from.
# iters will determine the length of the result text
generatedTokens = generate(model, n, seed, iters)
print "".join(generatedTokens)