Datasets:
Tasks:
Token Classification
Modalities:
Text
Sub-tasks:
sentiment-classification
Languages:
Polish
Size:
1K - 10K
License:
fix class distribution (#3)
Browse files- fix class distribution (750cf29db0831598ea066c31661757939415c636)
- fix class distribution in readme (bf59db862744fbb92408bc66cf45b692348203ad)
Co-authored-by: Albert Sawczyn <[email protected]>
README.md
CHANGED
@@ -55,17 +55,16 @@ Aspect-based sentiment analysis (ABSA) is a text analysis method that categorize
|
|
55 |
| val | 0 |
|
56 |
| test | 292 |
|
57 |
|
58 |
-
## Class distribution
|
59 |
-
|
60 |
-
| Class | train |
|
61 |
-
|
62 |
-
|
|
63 |
-
|
|
64 |
-
|
|
65 |
-
|
|
66 |
-
|
|
67 |
-
|
|
68 |
-
| a_amb | 0.0033 | - | 0.0024 |
|
69 |
|
70 |
## Citation
|
71 |
|
|
|
55 |
| val | 0 |
|
56 |
| test | 292 |
|
57 |
|
58 |
+
## Class distribution(without "O")
|
59 |
+
|
60 |
+
| Class | train | validation | test |
|
61 |
+
|:----------|--------:|-------------:|-------:|
|
62 |
+
| a_plus_m | 0.359 | - | 0.369 |
|
63 |
+
| a_minus_m | 0.305 | - | 0.377 |
|
64 |
+
| a_zero | 0.234 | - | 0.182 |
|
65 |
+
| a_minus_s | 0.037 | - | 0.024 |
|
66 |
+
| a_plus_s | 0.037 | - | 0.015 |
|
67 |
+
| a_amb | 0.027 | - | 0.033 |
|
|
|
68 |
|
69 |
## Citation
|
70 |
|