Numerical Claim Detection in Finance: A New Financial Dataset, Weak-Supervision Model, and Market Analysis

Battle Count: 135

The paper introduces a novel measure of optimism derived from claim detection, which shows potential for predicting earnings surprises and cumulative abnormal returns. This could be valuable for developing quantitative trading strategies.

The weak-supervision model and fine-tuned language models require significant NLP expertise to implement. However, the provided code and datasets should facilitate easier implementation for practitioners.

The paper provides detailed information on the methodology, datasets, and model implementations. Code and datasets will be made publicly available on GitHub and Hugging Face.