摘要
本文运用人工神经网络方法,研究互联网搜索对期权隐含波动率的影响.基于标普500指数看涨期权,本文首先建立了一个揭示期权隐含波动率变化与指数收益率、期权Delta和期权剩余有效期之间关系的人工神经网络模型,结果显示此模型的估计精度比Hull和White(2017)提出的解析模型提升了约15%,比Cao、Chen和Hull(2020)提出的神经网络模型提升了约40%.接着,本文引入25个互联网搜索关注度指标,将它们集成一个谷歌趋势指数并作为度量互联网搜索关注度的综合指标.最后,将该谷歌趋势指数的变化率加入到前述人工神经网络模型,从互联网搜索关注度的角度探究期权隐含波动率的动态特征,新的模型进一步提升了约30%的估计精度.
In this paper,we use the artificial neural network(ANN)method to investigate the effect of internet searches on the implied volatility of options.Based on the S&P 500 index call options,we establish an ANN model that reveals the relationship between the change of option implied volatility and the index return,option delta,and option maturity.The results show that the estimation accuracy of this model is 15%higher than that of the analytical model proposed by Hull and White(2017).This model is also 40%better than the neural network model of Cao,Chen,and Hull(2020).Then,we introduce 25 indicators of internet search attention,integrate them into a Google trends index(GTX),and use it as a comprehensive index to measure internet search attention.Finally,the change rate of the GTX is added to the ANN model mentioned above to explore the dynamic characteristics of option implied volatility from the perspective of internet search attention.The new model further improves the estimation accuracy by 30%.
作者
李星毅
刘彦初
朱书尚
LI Xingyi;LIU Yanchu;ZHU Shushang(School of Business,Sun Yat-sen University,Guangzhou 510275,China;Department of Management,Technology,and Economics,Eidgenössische Technische Hochschule Zürich,Zürich 8092,Switzerland;Lingnan College,Sun Yat-sen University,Guangzhou 510275,China)
出处
《系统工程理论与实践》
EI
CSSCI
CSCD
北大核心
2023年第7期2055-2071,共17页
Systems Engineering-Theory & Practice
基金
国家自然科学基金重大项目(71991474)
国家自然科学基金创新研究群体项目(71721001)
国家自然科学基金面上项目(72271249)
国家留学基金(202106380104)。
