---
license: mit
task_categories:
- text-classification
language:
- en
size_categories:
- 1K<n<10K
---
This dataset contains two manually pre-labeled datasets:

In the **economic agents dataset**, we labeled 6,205 randomized sentences from a [Fed database](https://github.com/Moritz-Pfeifer/CentralBankRoBERTa/tree/main/Data/FED) containing speeches (1948-2023) as speaking either about households, firms, the financial sector, the government, or the central bank itself. 

In the **sentiment dataset**, we labeled 6,683 randomized sentences from the same database, which are either labeled as being positive (1) or negative (0).

The datasets were used to train an [agent classifier](https://huggingface.co/Moritz-Pfeifer/CentralBankRoBERTa-agent-classifier) and a [sentiment classifier](https://huggingface.co/Moritz-Pfeifer/CentralBankRoBERTa-sentiment-classifier). 

<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". <em>Journal of Finance and Data Science </em> (forthcoming) <a href="https://doi.org/10.1016/j.jfds.2023.100114">https://doi.org/10.1016/j.jfds.2023.100114</a> 
    </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>