A0815
Title: Qualitative insights meet quantitative forecasts with GPT-4o mini for Taiwan earnings prediction
Authors: Yu Chuan Shih - National Yang Ming Chiao Tung University (Taiwan)
Hsin-Pei Huang - National Yang Ming Chiao Tung University (Taiwan)
Huei-Wen Teng - National Yang Ming Chiao Tung University (Taiwan)
Ming-Hsuan Kang - National Yang Ming Chiao Tung University (Taiwan) [presenting]
Abstract: The potential of large language models (LLMs) is explored, particularly GPT-4o mini, to enhance earnings forecasting by transforming structured financial data into qualitative textual analysis. Using income statements, balance sheets, and statements of cash flows from Taiwan-listed firms, GPT-4o mini generates descriptive reports through carefully designed prompts, including Chain-of-Thought (CoT) reasoning. These textual outputs are converted into numerical embeddings, which are subsequently used in prediction models. GPT-4o minis forecasts are benchmarked against traditional classifiers such as logistic regression and XGBoost. Results reveal that while GPT-4o mini alone does not outperform established methods, CoT prompting significantly improves its predictive performance and interpretability. The complementary role that LLMs can play in financial analysis is highlighted, bridging qualitative narrative understanding and quantitative forecasting.
