英文论文


文献类型
Article
题名
Deep learning, textual sentiment, and financial market
作者
Jiang, Fuwei; Liu, Yumin; Meng, Lingchao; Zhang, Huajing
作者单位
[Jiang, Fuwei; Liu, Yumin] Cent Univ Finance & Econ, Sch Finance, Beijing, Peoples R China. [Jiang, Fuwei] Xiamen Univ, Sch Econ, Xiamen, Peoples R China. [Jiang, Fuwei] Xiamen Univ, Wang Yanan Inst Studies Econ WISE, Xiamen, Peoples R China. [Meng, Lingchao] Univ Int Business & Econ, China Sch Banking & Finance, Beijing, Peoples R China. [Zhang, Huajing] Shandong Technol & Business Univ, Sch Finance, Yantai, Peoples R China.
通讯作者地址
Cent Univ Finance & Econ, Sch Finance, Beijing, Peoples R China.; Xiamen Univ, Sch Econ, Xiamen, Peoples R China.; Xiamen Univ, Wang Yanan Inst Studies Econ WISE, Xiamen, Peoples R China.; Shandong Technol & Business Univ, Sch Finance, Yantai, Peoples R China.
Email
jfuwei@gmail.com; yuminliu123@163.com; lingchao_meng@pku.edu.cn; hjingzhang@126.com
ResearchID
ORCID
期刊名称
INFORMATION TECHNOLOGY & MANAGEMENT
出版社
SPRINGER
ISSN
1385-951X
出版信息
2025-12, 26 (4):441-465.
JCR
影响因子
ISBN
基金
National Social Science Fund of China [22ZD063]; National Natural Science Foundation of China [72072193,71872195, 72342019]; Program for Innovation Research in CUFE
会议名称
会议地点
会议开始日期
会议结束日期
关键词
Textual sentiment; Deep learning; Sentiment dictionary; Asset pricing
摘要
In this paper, we apply the BERT model, a cut-edging deep learning model, to construct a novel textual sentiment index in the Chinese stock market. By introducing the stock market returns as sentiment labels, our BERT model effectively extracts textual sentiment-related information useful for asset pricing. We find that the BERT-based sentiment has much greater predictive power for stock market returns than the traditional dictionary method as well as the Baker-Wurgler investor sentiment index both in and out of sample. The BERT-based sentiment shows strong predictive power during economic downturns and can significantly predict future macroeconomic conditions. Overall, our BERT model offers a better measure of textual investor sentiment, highlighting the potentially significant value of deep learning, AI, and FinTech in financial market.
一级学科
Information Science & Library Science; Management
WOS入藏号
WOS:001589287600001
EI收录号
DOI
10.1007/s10799-024-00428-z
ESI
收录于
SSCI

返回