Chinese Fine-Grained Financial Sentiment Analysis with Large Language Models

• DOI: 10.48550/arxiv.2306.14096
• Publication Date: 2023-01-01
• Authors: Yinyu Lan; Yanru Wu; Wang Xu; Weiqiang Feng; Youhao Zhang
• Keywords: Sentiment Analysis

Entity-level fine-grained sentiment analysis in the financial domain is a crucial subtask of and currently faces numerous challenges. The primary challenge stems from lack high-quality large-scale annotated corpora specifically designed for text analysis, which turn limits availability data necessary developing effective processing techniques. Recent advancements large language models (LLMs) have yielded remarkable performance natural tasks, primarily centered around pattern matching. In this paper, we propose novel extensive Chinese dataset, FinChina SA, enterprise early warning. We thoroughly evaluate experiment with well-known existing open-source LLMs using our dataset. firmly believe that dataset will serve as valuable resource to advance exploration real-world should be focus future research. SA publicly available at https://github.com/YerayL/FinChina-SA