SILC-EFSA: Self-aware In-context Learning Correction for Entity-level Financial Sentiment Analysis

TL;DR

This paper introduces SILC-EFSA, a novel two-stage approach utilizing large language models and GNN-based correction to improve entity-level financial sentiment analysis, supported by the largest English and Chinese datasets.

Contribution

The paper constructs the largest entity-level financial sentiment datasets and proposes a two-stage SILC method combining LLM fine-tuning and GNN correction for improved accuracy.

Findings

Achieved state-of-the-art results on new datasets.

Demonstrated practical utility in cryptocurrency market monitoring.

Provided publicly available datasets and code.

Abstract

In recent years, fine-grained sentiment analysis in finance has gained significant attention, but the scarcity of entity-level datasets remains a key challenge. To address this, we have constructed the largest English and Chinese financial entity-level sentiment analysis datasets to date. Building on this foundation, we propose a novel two-stage sentiment analysis approach called Self-aware In-context Learning Correction (SILC). The first stage involves fine-tuning a base large language model to generate pseudo-labeled data specific to our task. In the second stage, we train a correction model using a GNN-based example retriever, which is informed by the pseudo-labeled data. This two-stage strategy has allowed us to achieve state-of-the-art performance on the newly constructed datasets, advancing the field of financial sentiment analysis. In a case study, we demonstrate the enhanced practical utility of our data and methods in monitoring the cryptocurrency market. Our datasets and code are available at https://github.com/NLP-Bin/SILC-EFSA.