Evaluating LLM Sentiment Factors for Chinese Stock Trading

Summary

This paper outlines an experimental framework for testing whether large language models can extract useful sentiment signals from Chinese financial news. It applies three different models, selected to represent distinct approaches to improving model performance, to company related news summaries. The resulting sentiment measures are treated as candidate quantitative factors, linking text analysis to portfolio decisions rather than evaluating language output alone.

The study then builds trading strategies from those factors and assesses them through backtests intended to reflect realistic trading conditions. Its central contribution, as described here, is a standardized procedure for comparing models and their downstream trading use. The excerpt gives no model names, backtest period, benchmark, numerical results, or detailed safeguards against leakage and transaction costs, so it does not establish which model or strategy performs best. The findings are framed as an evaluation of potential rather than proof of durable predictive power.

Key ideas

• LLMs can be evaluated for sentiment extraction from Chinese financial news summaries.
• The study compares three models that represent different performance enhancement approaches.
• Extracted sentiment is used to form quantitative factors and trading strategies.
• Backtesting connects language model outputs to investment performance assessment.
• The excerpt provides no numerical results or details needed to judge robustness and predictive durability.

Tags

# Unveiling the Potential of Sentiment: Can Large Language Models Predict Chinese Stock Price Movements?


# Unveiling the Potential of Sentiment: Can Large Language Models Predict Chinese Stock Price Movements?









The rapid advancement of Large Language Models (LLMs) has spurred discussions about their potential to enhance quantitative trading strategies. LLMs excel in analyzing sentiments about listed companies from financial news, providing critical insights for trading decisions. However, the performance of LLMs in this task varies substantially due to their inherent characteristics. This paper introduces a standardized experimental procedure for comprehensive evaluations. We detail the methodology using three distinct LLMs, each embodying a unique approach to performance enhancement, applied specifically to the task of sentiment factor extraction from large volumes of Chinese news summaries. Subsequently, we develop quantitative trading strategies using these sentiment factors and conduct back-tests in realistic scenarios. Our results will offer perspectives about the performances of Large Language Models applied to extracting sentiments from Chinese news texts.

Shown in full with attribution under the source's licence. Licence: abstract CC0

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.