search

posterior / distributions

Low-level primitives for collapsed Gibbs sampling in python and C++
BSD 3-Clause "New" or "Revised" License
16 stars 8 forks source link

discrete_goodness_of_fit bounds error #1

Closed fritzo closed 9 years ago

fritzo commented 9 years ago

distributions/util.py line 113

for sample in samples:
    counts[min(int(bin_count * sample), bin_count - 1)] += 1
return multinomial_goodness_of_fit(probs, counts, len(samples), plot=plot)