Not All News Is Equal: Topic- and Event-Conditional Sentiment from Finetuned LLMs for Aluminum Price Forecasting

2026-03-09

AI Summary

This study investigates how sentiment signals extracted from news texts can enhance the accuracy of aluminum price forecasting, particularly in highly volatile market conditions. By fine-tuning the Qwen3 large language model, the authors derive topic- and event-conditioned sentiment scores from English and Chinese news headlines, which are then integrated with structured features such as macroeconomic indicators and metal indices into an LSTM-based predictive model. The research systematically demonstrates, for the first time, the differential contributions of news sources, topics, and event types to aluminum price prediction. Empirical results on Shanghai Metal Exchange data from 2007 to 2024 show that the sentiment-augmented model achieves a Sharpe ratio of 1.04 during high-volatility periods, substantially outperforming a baseline model using only structured data (Sharpe ratio: 0.23), thereby confirming the significant economic value of sentiment signals generated by fine-tuned LLMs.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

By capturing the prevailing sentiment and market mood, textual data has become increasingly vital for forecasting commodity prices, particularly in metal markets. However, the effectiveness of lightweight, finetuned large language models (LLMs) in extracting predictive signals for aluminum prices, and the specific market conditions under which these signals are most informative, remains under-explored. This study generates monthly sentiment scores from English and Chinese news headlines (Reuters, Dow Jones Newswires, and China News Service) and integrates them with traditional tabular data, including base metal indices, exchange rates, inflation rates, and energy prices. We evaluate the predictive performance and economic utility of these models through long-short simulations on the Shanghai Metal Exchange from 2007 to 2024. Our results demonstrate that during periods of high volatility, Long Short-Term Memory (LSTM) models incorporating sentiment data from a finetuned Qwen3 model (Sharpe ratio 1.04) significantly outperform baseline models using tabular data alone (Sharpe ratio 0.23). Subsequent analysis elucidates the nuanced roles of news sources, topics, and event types in aluminum price forecasting.

Problem

• aluminum price forecasting
• sentiment analysis
• news topics
• market events
• commodity markets

Innovation

• finetuned LLMs
• topic-conditional sentiment
• aluminum price forecasting
• multilingual news analysis
• hybrid predictive modeling