Skip to content
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

[FIX] ScoringSheetViewer computing target class values from base values #7025

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

[FIX] ScoringSheetViewer computing target class values from base values #7025

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
43 changes: 34 additions & 9 deletions Orange/widgets/visualize/owscoringsheetviewer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -338,8 +338,11 @@ def __init__(self):
self.instance = None
self.instance_points = []
self.classifier = None
self.coefficients = None
self._base_coefficients = None
self._base_all_scores = None
self._base_all_risks = None
self.attributes = None
self.coefficients = None
self.all_scores = None
self.all_risks = None
self.domain = None
Expand Down Expand Up @@ -463,14 +466,16 @@ def _adjust_for_target_class(self):
This allows user to select the target class and see the
corresponding coefficients, scores, and risks.
"""
# Negate the coefficients
self.coefficients = [-coef for coef in self.coefficients]
# Negate the scores
self.all_scores = [-score if score != 0 else score for score in self.all_scores]
self.all_scores.sort()
# Adjust the risks
self.all_risks = [100 - risk for risk in self.all_risks]
self.all_risks.sort()
if self.target_class_index == 0:
self.coefficients = self._base_coefficients[:]
self.all_scores = self._base_all_scores[:]
self.all_risks = self._base_all_risks[:]
else:
self.coefficients = [-coef for coef in self._base_coefficients]
self.all_scores = sorted(
[-score if score != 0 else score for score in self._base_all_scores]
)
self.all_risks = sorted([100 - risk for risk in self._base_all_risks])

# Classifier Input Methods ---------------------------------------------------------------------

Expand Down Expand Up @@ -510,15 +515,32 @@ def _extract_data_from_model(self, classifier):
all_risks = 1 / (1 + np.exp(-all_scaled_scores))

self.attributes = attributes
self._base_coefficients = coefficients
self._base_all_scores = all_scores.tolist()
self._base_all_risks = (all_risks * 100).tolist()
self.coefficients = coefficients
self.all_scores = all_scores.tolist()
self.all_risks = (all_risks * 100).tolist()
self.domain = classifier.domain

# For some reason when leading the model the scores and probabilities are
# set for the wrong target class. This is a workaround to fix that.
self._sort_attributes_by_importance()
self._adjust_for_target_class()

def _sort_attributes_by_importance(self):
"""
Sorts the attributes in descending order of importance (absolute coefficient value)
by combining the attributes and coefficients into tuples and sorting them.
"""
combined = list(
zip(self.attributes, self.coefficients, self._base_coefficients)
)
combined_sorted = sorted(combined, key=lambda t: abs(t[1]), reverse=True)
self.attributes, self.coefficients, self._base_coefficients = map(
list, zip(*combined_sorted)
)

def _is_valid_classifier(self, classifier):
"""Check if the classifier is a valid ScoringSheetModel."""
if not isinstance(classifier, ScoringSheetModel):
Expand All @@ -533,6 +555,9 @@ def _clear_classifier_data(self):
self.all_scores = None
self.all_risks = None
self.classifier = None
self._base_coefficients = None
self._base_all_scores = None
self._base_all_risks = None
self.Outputs.features.send(None)

# Data Input Methods ---------------------------------------------------------------------------
Expand Down
Loading