摘要
在近年来人工智能领域的迅猛发展中,机器学习算法在股票价格预测中的应用引起了广泛关注。机器学习的基本原理及其在股票市场中的应用被逐步深入研究。根据不同的算法模型,探讨了支持向量机(SVM)、回归树、神经网络和集成学习四种主要模型的应用。神经网络部分特别关注了前馈神经网络、循环神经网络、卷积神经网络及长短期记忆网络的特性和应用。在集成学习部分,重点分析了XGBoost算法的优势。通过利用贵州茅台股份(600519)的历史交易数据,包括交易日期、开盘价、收盘价、最低价、最高价、成交量、价格变动和涨跌幅等,对上述模型进行了训练和预测。预测结果显示,SVM和CNN模型在拟合度和误差方面表现突出,而综合考虑各项指标后,XGBoost模型在综合性能上较为优越,因此被认为是最合适的模型。
In recent years,the rapid development of artificial intelligence has led to widespread attention to the application of machine learning algorithms in stock price prediction.The fundamental principles of machine learning and its applications in the stock market have been gradually explored.This study discusses the application of four major models based on different algorithms:Support Vector Machine(SVM),Regression Trees,Neural Networks,and Ensemble Learning.The neural network section specifically focuses on the characteristics and applications of Feedforward Neural Networks,Recurrent Neural Networks,Convolutional Neural Networks,and Long Short-Term Memory Networks.The Ensemble Learning section highlights the advantages of the XGBoost algorithm.By using historical trading data of Kweichow Moutai Co.,Ltd.(600519),including trading date,opening price,closing price,lowest price,highest price,trading volume,price changes,and percentage change,the models were trained and predicted.The results show that the SVM and CNN models performed excellently in terms of fit and error,while considering all indicators,the XGBoost model demonstrated superior overall performance,making it the most suitable model.
作者
秦研
孔浩铭
QIN Yan;KONG Hao-ming(School of Economics,Nanjing University of Posts and Telecommunications,Jiangsu,Nanjing 210023;Department of Mathematics,National University of Singapore,Singapore 119077)
出处
《汕头大学学报(人文社会科学版)》
2025年第7期25-34,94,95,共12页
Journal of Shantou University(Humanities and Social Sciences Edition)
基金
国家社会科学基金重大项目“金融业制度型开放助推构建新发展格局的路径研究”(23&ZD059)。
