Authors: Olivier Gandouet, Mouloud Belbahri, Armelle Jezequel, Yuriy Bodjov

Published on: March 04, 2024

Impact Score: 7.4

Arxiv code: Arxiv:2403.02185

Summary

• What is new: The use of ChatGPT to create models that generate easily interpretable features for evaluating financial outcomes from earnings calls, combining knowledge distillation and transfer learning.
• Why this is important: The difficulty of generating interpretable features from earnings calls for financial outcome prediction.
• What the research proposes: Developing lightweight models via knowledge distillation and transfer learning that can classify topics and sentiments effectively.
• Results: The models showed high accuracy in classifying topics and sentiments from earnings calls and were successfully applied in quantitative investing scenarios.

Technical Details

Technological frameworks used: Knowledge distillation, transfer learning

Models used: ChatGPT

Data used: Dataset annotated by financial experts

Potential Impact

Financial analysis firms, investment firms, quantitative hedge funds

Want to implement this idea in a business?

We have generated a startup concept here: FinTrend.