随机森林与XGBoost在股票价格预测中的应用

Application of Random Forest and XGBoost in Stock Price Prediction

摘要: 金融市场具有强随机性、非线性、非平稳性等特征，对股票价格预测提出了挑战。本文引入支持向量机(SVM)、决策树、随机森林、XGBoost等四种机器学习算法，通过数据收集、数据预处理、模型训练、结果分析等步骤，对股票价格进行预测。实验结果表明，四类模型均能在一定程度上捕捉股价变化趋势；综合考虑预测效果和稳定性，随机森林和XGBoost模型更具优势，为预测股票价格提供了有益参考。

Abstract: Financial markets are characterized by high randomness, nonlinearity, and non-stationarity, posing significant challenges to stock price prediction. This study introduces four machine learning algorithms—Support Vector Machine (SVM), Decision Tree, Random Forest, and XGBoost—to predict stock prices through a process encompassing data collection, data preprocessing, model training, and result analysis. Experimental results demonstrate that all four models can, to some extent, capture the trends in stock price movements. When considering both predictive performance and stability, the Random Forest and XGBoost models exhibit superior advantages, offering valuable insights for stock price prediction.