高频指数强上涨趋势预测研究

Prediction of Significant up Trend for High Frequency Index

随着我国经济建设的发展,预测股票市场的走势已经是各投资机构和广大散户关注的一个焦点.由于股票市场波动受到各种因素的影响,具有极大的复杂性且频繁交易产生一定的费用,现有策略和模型不能很好地解决这个问题.利用深度学习知识,提出强上涨预测模型.将预测任务定义为强上涨、震荡和强下跌三分类问题,有效降低在预测为强上涨趋势时出现错误的概率,且减少在震荡阶段由于频繁交易产生的手续费.所提出的模型在沪深300指数5分钟数据集上性能优于其他基准模型,能帮助投资者获取稳定的收益,避免巨大亏损.

With the development of China's economic construction,predicting the trend of the stock market has become a focus of investment institu tions and the majority of retail investors.Because the stock market fluctuations are affected by various factors,which have great complexity and frequent transactions generate certain costs,existing strategies and models cannot solve this problem well.Uses deep learning knowl edge to propose a strong rising prediction model.Defining the prediction task as the three classification problems of strong rise,shock,and strong decline,effectively reducing the probability of errors when predicting a strong upward trend,and reducing the stock exchange fee due to frequent transactions during the shock phase.The proposed model outperforms other benchmark models on the dataset of 5 minute CSI 300 Index,which can help investors obtain stable returns and avoid huge losses.