2martens/masterthesis
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Made f1 score calculation robust against classes with zero predictions

Browse Source 
Signed-off-by: Jim Martens <github@2martens.de>
This commit is contained in:
Jim Martens
2019-06-10 16:49:17 +02:00
parent 15498549b7
commit 488703bb26
1 changed files with 3 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
src/twomartens/masterthesis/evaluate.py
View File

@ -293,6 +293,9 @@ def get_f1_score(cumulative_precisions: List[np.ndarray],
for class_id in range(1, nr_classes + 1): for class_id in range(1, nr_classes + 1):
cumulative_precision = cumulative_precisions[class_id] cumulative_precision = cumulative_precisions[class_id]
cumulative_recall = cumulative_recalls[class_id] cumulative_recall = cumulative_recalls[class_id]
if (cumulative_precision + cumulative_recall) == 0:
cumulative_f1_scores.append([])
continue
f1_score = 2 * ((cumulative_precision * cumulative_recall) / (cumulative_precision + cumulative_recall)) f1_score = 2 * ((cumulative_precision * cumulative_recall) / (cumulative_precision + cumulative_recall))
cumulative_f1_scores.append(f1_score) cumulative_f1_scores.append(f1_score)