基于卷积神经网络的股指期货套利策略研究

Research on Stock Index Futures Arbitrage Strategy Based on Convolutional Neural Network

摘要: 在人工智能技术迅猛发展的背景下，神经网络与金融领域的结合越来越密切，利用神经网络等机器学习方法构建投资策略，已是当前投资界和学术界的一个热点问题。提出了一种基于卷积神经网络的股指期货套利策略。通过建立卷积神经网络模型对股指期货价差进行预测，采用动态阈值方法确定套利区间。并对中证500股指期货和上证50股指期货进行了实证分析。对比研究基于SVM模型和XGboost模型的交易策略。实证结果显示，在最大回撤率处于同一水平下基于卷积神经网络的股指期货套利策略收益率比基于SVM模型和XGboost模型的套利策略收益率表现更好。证实了将卷积神经网络应用在股指期货套利策略当中具有一定的可行性与有效性。

Abstract: With the rapid development of artificial intelligence technology, the combination of neural network and financial field is closer. It is a hot issue in the field of investment and academia to construct investment strategy by using machine learning methods such as neural network. This paper proposes a stock index futures arbitrage strategy based on convolutional neural network. The convolution neural network model is established to predict the spread of stock index futures, and the dynamic threshold method is used to determine the arbitrage interval. This paper makes an empirical analysis on the CSI 500 stock index futures and the SSE 50 stock index futures. In this paper, we make a comparative study of the trading strategy based on SVM model and XGboost model. The empirical results show that when the maximum withdrawal rate is at the same level, the return rate of stock index futures arbitrage strategy based on convolutional neural network is better than that based on SVM model and XGboost model. It is proved that the convolution neural network is feasible and effective in the arbitrage strategy of stock index futures.