languages/cpython
languages/cpython
1
0
Fork 0
mirror of https://github.com/python/cpython synced 2024-09-04 15:56:13 +00:00
Code Issues Packages Projects Releases Wiki Activity

Improve accuracy for Spearman's rank correlation coefficient. (#96392)

Browse source
This commit is contained in:
Raymond Hettinger 2022-08-29 12:19:48 -05:00 committed by GitHub
parent 873554ef84
commit 3d180e3ab2
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 4 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

7
Lib/statistics.py
View file

@ -379,7 +379,7 @@ def _rank(data, /, *, key=None, reverse=False, ties='average', start=1) -> list[
[2.0, 1.0, 3.0]
Ranks are conventionally numbered starting from one; however,
setting *start* to zero allow the ranks to be used as array indices:
setting *start* to zero allows the ranks to be used as array indices:
>>> prize = ['Gold', 'Silver', 'Bronze', 'Certificate']
>>> scores = [8.1, 7.3, 9.4, 8.3]
@ -1073,8 +1073,9 @@ def correlation(x, y, /, *, method='linear'):
if method not in {'linear', 'ranked'}:
raise ValueError(f'Unknown method: {method!r}')
if method == 'ranked':
x = _rank(x)
y = _rank(y)
start = (n - 1) / -2 # Center rankings around zero
x = _rank(x, start=start)
y = _rank(y, start=start)
xbar = fsum(x) / n
ybar = fsum(y) / n
sxy = fsum((xi - xbar) * (yi - ybar) for xi, yi in zip(x, y))