异构数据融合驱动的股市波动预测研究  

Research on stock market volatility prediction driven by heterogeneous data fusion

作　　者：叶慕戎 鲁越 谭楚婷 YE Murong;LU Yue;TAN Chuting(Shanghai Lixin University of Accounting and Finance,Shanghai 201209,China)

机构地区：[1]上海立信会计金融学院,上海201209

出　　处：《计算机应用文摘》2023年第13期86-88,共3页Chinese Journal of Computer Application

基　　金：上海立信会计金融学院大学生创新创业训练计划项目(S202211047111).

摘　　要：由于股票市场具有复杂性、动态性和混乱性等诸多特点,其波动易受各种信息源的影响,因此对其预测具有相当的挑战性,而机器学习方法的应用在目前取得了一定的成功。文章从深度学习方法出发,融合多种数据源,提出一种异构数据融合驱动的神经网络模型,探索股市舆情、量化指标与股价波动的内在联系,以及媒体信息对股市波动的影响机制。Due to the complexity,dynamics and confusion of the stock market,its fluctuations are affected by various information sources,so its prediction is quite challenging,and the application of machine learning methods has achieved certain success at present.Starting from the deep learning method,integrating multiple data sources,this paper proposes a neural networks model driven by heterogeneous data fusion to explore the internal connection between stock market public opinion,quantitative indicators and stock price fluctuations,and the influence mechanism of media information on stock market fluctuations.

关 键 词：多特征融合 奥情分析 股市预测 LDA 神经网络 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]