DataWizardd commited on
Commit
27bafa1
Β·
verified Β·
1 Parent(s): 70420e5

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +20 -7
README.md CHANGED
@@ -16,12 +16,12 @@ model-index:
16
  type: text-classification
17
  metrics:
18
  - type: accuracy
19
- value: (직접 μΈ‘μ •ν•œ 정확도)
20
  ---
21
 
22
  # FinBERT Sentiment Analysis (Korean, Finance Domain)
23
 
24
- 이 λͺ¨λΈμ€ **ν•œκ΅­μ–΄ 금육 λ‰΄μŠ€ μš”μ•½λ¬Έ**을 λŒ€μƒμœΌλ‘œ 감정을 λΆ„λ₯˜ν•˜κΈ° μœ„ν•΄ νŒŒμΈνŠœλ‹λœ BERT 기반 λͺ¨λΈμž…λ‹ˆλ‹€.
25
  감정 λΆ„λ₯˜λŠ” λ‹€μŒ μ„Έ κ°€μ§€ 클래슀 쀑 ν•˜λ‚˜λ‘œ μˆ˜ν–‰λ©λ‹ˆλ‹€:
26
 
27
  - `0`: λΆ€μ •
@@ -31,13 +31,27 @@ model-index:
31
  ## 🧠 ν•™μŠ΅ 정보
32
 
33
  - 기반 λͺ¨λΈ: [`snunlp/KR-FinBERT-SC`](https://huggingface.co/snunlp/KR-FinBERT-SC)
34
- - 데이터: 직접 μˆ˜μ§‘ν•œ **넀이버 금육 λ‰΄μŠ€** μš”μ•½ + 감정 μˆ˜μž‘μ—… 라벨링
35
- - 총 μƒ˜ν”Œ 수: μ•½ 200개
36
  - Optimizer: AdamW
37
  - Epochs: 4
38
  - μ΅œλŒ€ 길이: 128
39
  - 평가 μ§€ν‘œ: Accuracy, F1 Score
40
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
41
  ## πŸ›  μ‚¬μš© 방법
42
 
43
  ```python
@@ -45,6 +59,5 @@ from transformers import pipeline
45
 
46
  pipe = pipeline("text-classification", model="DataWizardd/finbert-sentiment-ko")
47
 
48
- # μ˜ˆμ‹œ
49
- pipe("원/λ‹¬λŸ¬ ν™˜μœ¨μ΄ κΈ‰λ“±ν•˜λ©° μ‹œμž₯ λΆˆμ•ˆμ΄ μ»€μ‘Œλ‹€.")
50
- # 좜λ ₯: [{'label': 'LABEL_0', 'score': 0.95}] β†’ λΆ€μ •
 
16
  type: text-classification
17
  metrics:
18
  - type: accuracy
19
+ value: 0.93
20
  ---
21
 
22
  # FinBERT Sentiment Analysis (Korean, Finance Domain)
23
 
24
+ 이 λͺ¨λΈμ€ **ν•œκ΅­μ–΄ ν™˜μœ¨(금육) λ‰΄μŠ€ μš”μ•½λ¬Έ**을 λŒ€μƒμœΌλ‘œ 감정을 λΆ„λ₯˜ν•˜κΈ° μœ„ν•΄ νŒŒμΈνŠœλ‹λœ BERT 기반 λͺ¨λΈμž…λ‹ˆλ‹€.
25
  감정 λΆ„λ₯˜λŠ” λ‹€μŒ μ„Έ κ°€μ§€ 클래슀 쀑 ν•˜λ‚˜λ‘œ μˆ˜ν–‰λ©λ‹ˆλ‹€:
26
 
27
  - `0`: λΆ€μ •
 
31
  ## 🧠 ν•™μŠ΅ 정보
32
 
33
  - 기반 λͺ¨λΈ: [`snunlp/KR-FinBERT-SC`](https://huggingface.co/snunlp/KR-FinBERT-SC)
34
+ - 데이터: 직접 μˆ˜μ§‘ν•œ **넀이버 ν™˜μœ¨(금육) λ‰΄μŠ€** μš”μ•½ + 감정 μˆ˜μž‘μ—… 라벨링
35
+ - 총 μƒ˜ν”Œ 수: μ•½ 200
36
  - Optimizer: AdamW
37
  - Epochs: 4
38
  - μ΅œλŒ€ 길이: 128
39
  - 평가 μ§€ν‘œ: Accuracy, F1 Score
40
 
41
+ ## πŸ“Š μ„±λŠ₯ 평가
42
+
43
+ | 감정 클래슀 | Precision | Recall | F1-score | Support |
44
+ |-------------|-----------|--------|----------|---------|
45
+ | λΆ€μ • | 0.89 | 1.00 | 0.94 | 17 |
46
+ | 쀑립 | 1.00 | 0.82 | 0.90 | 11 |
47
+ | 긍정 | 0.93 | 0.93 | 0.93 | 14 |
48
+ | **정확도** | | | **0.93** | 42 |
49
+
50
+ > 전체 정확도: **93%**
51
+ > Macro F1-score: **0.92**
52
+
53
+ ---
54
+
55
  ## πŸ›  μ‚¬μš© 방법
56
 
57
  ```python
 
59
 
60
  pipe = pipeline("text-classification", model="DataWizardd/finbert-sentiment-ko")
61
 
62
+ pipe("ν™˜μœ¨μ΄ κΈ‰λ“±ν•˜λ©° μ‹œμž₯ λΆˆμ•ˆμ΄ 컀지고 μžˆλ‹€.")
63
+ # 좜λ ₯: [{'label': 'λΆ€μ •', 'score': 0.95}]