基于HP滤波的股票长期价格神经网络模型预测

Neural network model prediction of long-term stock prices based on HP filter

将上证指数和深证成指股票数据作为研究对象,对股票长期价格进行预测.选取长短期神经网络、循环神经网络、HP滤波长短期神经网络混合模型和HP滤波循环神经网络混合模型进行比较分析.经过模型间的对比分析,发现HP滤波对长短期神经网络预测的优化效果要优于循环神经网络.

Using the Shanghai composite index and Shenzhen composite index stock data as research objects,the long-term stock prices is predicted.Long and short term neural network,recurrent neural network,HP filtered long and short term neural network hybrid models and HP filtered recurrent neural network hybrid models are selected for comparative analysis.Through comparative analysis between models,it is found that HP filtering has better optimization effects on the long and short term neural network models than recurrent neural network models.