Skip to content

Fix calculation of agreement for π*, maybe. #3

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
rybesh wants to merge 3 commits into
base: main
Choose a base branch
from
Open

Fix calculation of agreement for π*, maybe. #3

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
5c7a47c
Fix calculation of agreement for π*, maybe.
rybesh Mar 12, 2015
c46e12d
We are counting (near) matches, not edits, so we want 1-weight.
rybesh Mar 25, 2015
7cd095c
Hypothesis is on the left (a), reference is on the right (b).
rybesh Mar 25, 2015
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 4 additions & 6 deletions segeval/agreement/pi.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,6 +6,7 @@
from __future__ import absolute_import, division
from decimal import Decimal
from segeval.agreement import __fnc_metric__, __actual_agreement_linear__
from itertools import chain


def __fleiss_pi_linear__(dataset, **kwargs):
Expand All @@ -29,13 +30,10 @@ def __fleiss_pi_linear__(dataset, **kwargs):
A_a = Decimal(sum(all_numerators)) / sum(all_denominators)
# Calculate Ae
p_e_segs = list()
for boundaries_info in coders_boundaries.values():
for item in boundaries_info:
boundaries, total_boundaries = item
p_e_seg = Decimal(boundaries) / total_boundaries
p_e_segs.append(p_e_seg)
boundary_ratios = chain.from_iterable(coders_boundaries.values())
b_placed, b_possible = map(sum, zip(*boundary_ratios))
# Calculate P_e_seg
P_e_seg = Decimal(sum(p_e_segs)) / len(p_e_segs)
P_e_seg = Decimal(b_placed) / b_possible
A_e = (P_e_seg ** 2)
# Calculate pi
pi = (A_a - A_e) / (Decimal('1') - A_e)
Expand Down
8 changes: 4 additions & 4 deletions segeval/similarity/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -103,7 +103,7 @@ def __boundary_confusion_matrix__(*args, **kwargs):
# Add weighted near misses
for transposition in statistics['transpositions']:
match = transposition[2]
matrix[match][match] += fnc_weight_t([transposition], n_t)
matrix[match][match] += (1 - fnc_weight_t([transposition], n_t))
# Add confusion errors
for substitution in statistics['substitutions']:
hyp, ref = substitution
Expand All @@ -113,11 +113,11 @@ def __boundary_confusion_matrix__(*args, **kwargs):
hyp, ref = None, None
boundary_type, side = addition
if side == 'a':
hyp = None
ref = boundary_type
else: # side == 'b'
hyp = boundary_type
ref = None
else: # side == 'b'
hyp = None
ref = boundary_type
assert side == 'a' or side == 'b'
matrix[hyp][ref] += 1
return matrix
Expand Down