期权价格预测的LSTM网络  

LSTM Network for Option Price Prediction

作　　者：肖鑫 XIAO Xin(School of Economics and Management,Southeast University,Nanjing,China,211189)

机构地区：[1]东南大学经济管理学院,南京211189

出　　处：《福建电脑》2023年第12期20-23,共4页Journal of Fujian Computer

摘　　要：期权是重要的金融衍生品和风险管理工具之一,提高期权价格预测的精度对投资者的风险管理十分必要。本文提出了基于多源信息的LSTM模型来预测期权价格。首先通过金融数据库计算得到多源信息数据,然后构建单层LSTM网络在训练集上训练模型。用模型对测试集进行预测,得到均方根误差为0.0244619,平均绝对误差为0.0125933,R-squared为0.9616692。实证结果表明,模型具有较好的预测效果和泛化能力。Options are one of the important financial derivatives and risk management tools,and improving the accuracy of option price prediction is essential for investors'risk management.This article proposes an LSTM model based on multi-source information to predict option prices.Firstly,multi-source information data is calculated through financial databases,and then a single-layer LSTM network is constructed to train the model on the training set.Using the model to predict the test set,the root mean square error was 0.0244619,the average absolute error was 0.0125933,and the R-squared was 0.9616692.The empirical results indicate that the model has good predictive performance and generalization ability.

关 键 词：期权价格 长短期记忆神经网络 多源数据 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]