ahmedrachid
commited on
Commit
•
a1321da
1
Parent(s):
42d4e9f
Update README.md
Browse files
README.md
CHANGED
@@ -19,6 +19,25 @@ Our model was fine-tuned for Sentiment Analysis task on _FinancialPhraseBank_ da
|
|
19 |
### Training data
|
20 |
FinancialBERT model was fine-tuned on [Financial PhraseBank](https://www.researchgate.net/publication/251231364_FinancialPhraseBank-v10), a dataset consisting of 4840 Financial News categorised by sentiment (negative, neutral, positive).
|
21 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
22 |
### How to use
|
23 |
Our model can be used thanks to Transformers pipeline for sentiment analysis.
|
24 |
```python
|
|
|
19 |
### Training data
|
20 |
FinancialBERT model was fine-tuned on [Financial PhraseBank](https://www.researchgate.net/publication/251231364_FinancialPhraseBank-v10), a dataset consisting of 4840 Financial News categorised by sentiment (negative, neutral, positive).
|
21 |
|
22 |
+
### Fine-tuning hyper-parameters
|
23 |
+
- learning_rate = 2e-5
|
24 |
+
- batch_size = 32
|
25 |
+
- max_seq_length = 512
|
26 |
+
- num_train_epochs = 5
|
27 |
+
|
28 |
+
### Metrics
|
29 |
+
The evaluation metrics used are: Precision, Recall and F1-score. The following is the classification report on the test set.
|
30 |
+
| relation | precision | recall | f1-score | support |
|
31 |
+
| ------------- |:-------------:|:-------------:|:-------------:| -----:|
|
32 |
+
| has | 0.7416 | 0.9674 | 0.8396 | 2362 |
|
33 |
+
| is in | 0.7813 | 0.7925 | 0.7869 | 2362 |
|
34 |
+
| is | 0.8650 | 0.6863 | 0.7653 | 2362 |
|
35 |
+
| are | 0.8365 | 0.8493 | 0.8429 | 2362 |
|
36 |
+
| x | 0.9515 | 0.8302 | 0.8867 | 2362 |
|
37 |
+
| | | | | |
|
38 |
+
| macro avg | 0.8352 | 0.8251 | 0.8243 | 11810 |
|
39 |
+
| weighted avg | 0.8352 | 0.8251 | 0.8243 | 11810 |
|
40 |
+
|
41 |
### How to use
|
42 |
Our model can be used thanks to Transformers pipeline for sentiment analysis.
|
43 |
```python
|