pywsi.stats package¶

Module contents¶

pywsi.stats.welford_simulatenous_update(prev_avg, prev_M2, new_value, count)[source]¶

Perform Welford’s simulatenous update on mean and variance

M2 = n*sigma_n^2

Parameters:	prev_avg: array_like Vector of (count-1)th step averages prev_M2: array_like Vector of (count-1)th step M2 stats new_value: array_like New incoming values vector count: int Current count (count starts from 1)
Returns:	new_avg: array_like Updated average new_M2: array_like Updated M2 stats new_var: array_like Population variance new_samplevar: array_like Sample Variance

pywsi.stats.welford_update_M1(prev_avg, new_value, count)[source]¶

Welford’s updated mean.

mu_{n} = mu_{n-1} + 1/n(x_n - mu_{n-1}).

Parameters:	prev_avg: array_like Vector of (count-1)th step averages new_value: array_like New incoming values vector count: int Current count (count starts from 1)
Returns:	new_avg: array_like Updated average

pywsi.stats.welford_update_M2(prev_M2, new_value, prev_avg, new_avg)[source]¶

Welford’s updated variance.

S_n = S_{n-1} + (x_n-mu_{n-1})(x_n-mu_n)

Parameters:	prev_M2: array_like Vector of (count-1)th step M2 stats new_value: array_like New incoming values vector prev_avg: array_like Vector of (count-1)th step averages count: int Current count (count starts from 1)
Returns:	new_M2: array_like Updated M2 stats