多种机器学习方法在股票预测中的优势对比

Comparison of Advantages of Multiple Machine Learning Methods in Stock Prediction

摘要: 股票价格的预测是金融领域的一个重要研究课题，随着机器学习技术的发展，各种模型在股票预测中的应用也越来越广泛。本文通过比较BP神经网络(BP)、极限学习机(ELM)和长短期记忆网络(LSTM)三种常见机器学习方法的表现，分析其在股票预测中的优劣。通过对比实验数据的均方误差(MSE)、均方根误差(RMSE)、平均绝对误差(MAE)、和决定系数(R²)等评价指标，本文验证了LSTM在处理时间序列数据方面的优势，特别是在股票价格预测的任务中表现出色，而BP神经网络和ELM各自具有特定场景下的应用价值。

Abstract: Predicting stock prices is an important research topic in the field of finance, and with the development of machine learning technology, various models are increasingly being applied in stock price prediction. This paper compares the performance of three common machine learning methods, BP neural network (BP), extreme learning machine (ELM), and long short-term memory network (LSTM), to analyze their advantages and disadvantages in stock price prediction. By comparing the evaluation indicators such as mean square error (MSE), root mean square error (RMSE), average absolute error (MAE), and coefficient of determination (R²) of the experimental data, this paper verifies the advantage of LSTM in handling time series data, especially in the task of stock price prediction, and the excellent performance of LSTM in this task. Meanwhile, BP neural network and ELM each have their own application value in specific scenarios.