search

cgevans / scikits-bootstrap

Python/numpy bootstrap confidence interval estimation.
Other
174 stars 36 forks source link

Possible bug #12

Closed ottokart closed 7 years ago

ottokart commented 7 years ago

Compared to https://ocw.mit.edu/courses/mathematics/18-05-introduction-to-probability-and-statistics-spring-2014/readings/MIT18_05S14_Reading24.pdf you have reversed the deltas for upper and lower bound confidence intervals. One of you is probably wrong, but I'm new to bootstrapping, so it might be me who is wrong.

Also, instead of nvals = np.round((n_samples-1)avals), nvals = np.round(n_samplesavals)-1 should be better (gets integer without rounding for default alphas and n_samples)

ottokart commented 7 years ago

My bad, didn't notice you use percentile bootstrap instead of empirical bootstrap.