北京师范大学学报（自然科学版）2025，Vol.61Issue(6)：797-804,8.DOI:10.12202/j.0476-0301.2025141
融合大语言模型的金融多元时序预测框架及试验评估
A financial multivariate time series forecasting framework incorporating large language models and experimental evaluation
摘要
Abstract
Mainstream time series forecasting methods based on large language model(LLM)are analyzed,and a unified model framework is proposed to evaluate empirically exchange rate and stock index data.LLM exhibits certain performance advantages in financial time series forecasting,with certain notable limitations.Relying solely on simple textual inputs or prompts may not lead to performance improvements.关键词
金融时间序列/大语言模型/预测模型/金融市场Key words
financial time series/large language model/forecasting model/financial market分类
管理科学引用本文复制引用
肖雯艺琳,侯朝川,韩松乔..融合大语言模型的金融多元时序预测框架及试验评估[J].北京师范大学学报（自然科学版）,2025,61(6):797-804,8.基金项目
国家自然科学基金资助项目(72342009) （72342009）
