|
--- |
|
license: mit |
|
|
|
widget: |
|
- text: "The early effects of our policy tightening are also becoming visible, especially in sectors like manufacturing and construction that are more sensitive to interest rate changes." |
|
--- |
|
<div style="display: flex; align-items: center;"> |
|
<img src="https://i.postimg.cc/HLqPqkyk/Central-Bank-Ro-BERTa-logos-black.png" width="200" height="200" style="margin-right: 20px;"> |
|
<div> |
|
<h1 style="font-size: 36px; font-weight: bold; margin: 0;">CentralBankRoBERTa</h1> |
|
<p style="font-size: 18px; margin: 0;">A Fine-Tuned Large Language Model for Central Bank Communications</p> |
|
</div> |
|
</div> |
|
|
|
## CentralBankRoBERTa |
|
|
|
CentralBankRoBERTA is a large language model. It combines an economic [agent classifier](Moritz-Pfeifer/CentralBankRoBERTa-audience-classifier) that distinguishes five basic macroeconomic agents with a binary sentiment classifier that identifies the emotional content of sentences in central bank communications. |
|
|
|
#### Overview |
|
|
|
The SentimentClassifier model is designed to detect whether a given sentence is positive or negative for either **households**, **firms**, **the financial sector** or **the government**. This model is based on the RoBERTa architecture and has been fine-tuned on a diverse and extensive dataset to provide accurate predictions. |
|
|
|
#### Intended Use |
|
|
|
The AudienceClassifier model is intended to be used for the analysis of central bank communications where sentiment analysis is essential. |
|
|
|
#### Performance |
|
|
|
- Accuracy: 88% |
|
- F1 Score: 0.88 |
|
- Precision: 0.88 |
|
- Recall: 0.88 |
|
|
|
### Usage |
|
|
|
You can use these models in your own applications by leveraging the Hugging Face Transformers library. Below is a Python code snippet demonstrating how to load and use the AudienceClassifier model: |
|
|
|
```python |
|
from transformers import pipeline |
|
|
|
# Load the AudienceClassifier model |
|
audience_classifier = pipeline("text-classification", model="Moritz-Pfeifer/CentralBankRoBERTa-sentiment-classifier") |
|
|
|
# Perform audience classification |
|
sentinement_result = audience_classifier("The early effects of our policy tightening are also becoming visible, especially in sectors like manufacturing and construction that are more sensitive to interest rate changes.") |
|
print("Sentiment:", sentinement_result[0]['label']) |
|
``` |
|
|
|
<table> |
|
<tr> |
|
<td colspan="2" style="border-top: 1px solid #ccc; padding: 5px; text-align: left;"> |
|
Please cite this model as Pfeifer, M. and Marohl, V.P. (2023) "CentralBankRoBERTa: A Fine-Tuned Large Language Model for Central Bank Communications" ADD SOURCE/LINK |
|
</td> |
|
</tr> |
|
<tr> |
|
<td style="padding: 5px;"> |
|
Moritz Pfeifer<br> |
|
Institute for Economic Policy, University of Leipzig<br> |
|
04109 Leipzig, Germany<br> |
|
<a href="mailto:pfeifer@wifa.uni-leipzig.de">pfeifer@wifa.uni-leipzig.de</a> |
|
</td> |
|
<td style="padding: 5px;"> |
|
Vincent P. Marohl<br> |
|
Department of Mathematics, Columbia University<br> |
|
New York NY 10027, USA<br> |
|
<a href="mailto:vincent.marohl@columbia.edu">vincent.marohl@columbia.edu</a> |
|
</td> |
|
</tr> |
|
</table> |
|
|
