基于GRU神经网络研究不同证券市场对股票收益的影响——以恒生和上证指数为例

Study on the influence of different Securities Market on Stock return based on GRU Neural Network——Taking Hang Sheng and Shanghai Stock Exchange Index as an example

大量研究表明,不同金融市场之间具有不可忽视的联系,股票作为金融资本市场中最具代表性的一部分,与其他金融证券市场的联系更为紧密。因此,针对不同金融市场的相互影响问题,以恒生、上证指数为例,提出建立神经网络模型。以两种指数的每日数据为样本,利用GRU(递归神经网络)神经网络的时间记忆性能,刻画出在加入不同证券指数特征的影响下,对上证股票指数收益的波动情况进行预测研究。训练和测试结果表明,GRU神经网络模型效果较为理想,而加入恒生指数特征的预测效果最好。这可以为后续中外金融市场关系的研究提供一定的参考价值,对想要购买企业债券的操作者也具有较高的实际价值。

A large number of studies show that there is a relationship between different financial markets that can not be ignored,and stocks,as the most representative part of the financial capital market,are more closely related to other financial securities markets.Therefore,in view of the interaction between different financial markets,taking Hang Seng and Shanghai Stock Exchange Index as an example,a neural network model is proposed.Taking the daily data of the two indices as samples,using the time memory performance of GRU(recurrent neural network)neural network,this paper describes the volatility of Shanghai stock index return under the influence of different securities index characteristics.The training and test results show that the GRU neural network model is more effective,while the Hang Seng index feature is the best.This can provide a certain reference value for the follow-up study of the relationship between Chinese and foreign financial markets,and also has a high practical value for operators who want to buy corporate bonds.