【摘要】 高频金融数据和金融资产收益率是金融计量学的一个全新的研究领域。目前,国内学者利用年、月、日等低频数据对股票市场的收益率进行了很多的研究,但是以日内高频数据为基础的研究还不多见。如何较准确地预测基于高频数据的股票收益率是进一步深入研究金融市场的基础,论文采用数据挖掘中的BP神经网络对沪深300指数高频数据中的日内收益率进行建模与预测。结果表明:神经网络模型对股票高频数据的日内收益率具有很强的预测能力。更多还原

【Abstract】 The study of high -frequency financial data and stock market returns is a brand new field in financial econometrics, however, current studies on financial returns usually use low-frequency data rather than high-frequency data. How to better measure stock market returns based on high-frequency data is the basis of further study in the financial market. This paper attempts to build up a BP neural net by using high-frequency data of the Shanghai and Shenzhen 300 Index to forecast the returns. The results show that neural net is capable of measuring the returns of stock market based on high-frequency data.更多还原