在股票市场中,准确的股价预测对于投资者的决策制定和金融机构的风险管理具有重要意义。随着计算机技术的迅猛发展,越来越多的金融学者开始关注神经网络,利用网络技术建立股票收盘价趋势预测模型,取得了显著效果。本文基于matlab选择门...在股票市场中,准确的股价预测对于投资者的决策制定和金融机构的风险管理具有重要意义。随着计算机技术的迅猛发展,越来越多的金融学者开始关注神经网络,利用网络技术建立股票收盘价趋势预测模型,取得了显著效果。本文基于matlab选择门控循环单元模型探究股价预测。考虑到神经网络参数选择对于预测性能具有较大影响,因此利用改进的河马优化算法优化GRU模型参数,构建IHO-GRU模型对中信银行的收盘价进行预测,实验结果显示,IHO-GRU模型在RMSE、MAE、MAPE三个评价指标的表现均为最优,展示了优化算法和神经网络在金融市场预测中的潜力,也为实际金融投资提供了一定的参考性。Accurate stock price prediction is of great significance for investor decision-making and financial institution risk management in the stock market. With the rapid development of computer technology, more and more financial scholars are paying attention to neural networks and using network technology to establish a stock closing price trend prediction model, which has achieved significant results. This article explores stock price prediction based on the Gated Recurrent Unit model selected in Matlab. Considering that the selection of neural network parameters has a significant impact on predictive performance, an improved hippopotamus optimization algorithm was used to optimize the GRU model parameters and construct an IHO-GRU model to predict the closing price of CITIC Bank. The experimental results showed that the IHO-GRU model performed the best in the RMSE, MAE, and MAPE evaluation indicators, demonstrating the potential of optimization algorithms and neural networks in financial market prediction, and providing some reference for actual financial investment.展开更多
文摘在股票市场中,准确的股价预测对于投资者的决策制定和金融机构的风险管理具有重要意义。随着计算机技术的迅猛发展,越来越多的金融学者开始关注神经网络,利用网络技术建立股票收盘价趋势预测模型,取得了显著效果。本文基于matlab选择门控循环单元模型探究股价预测。考虑到神经网络参数选择对于预测性能具有较大影响,因此利用改进的河马优化算法优化GRU模型参数,构建IHO-GRU模型对中信银行的收盘价进行预测,实验结果显示,IHO-GRU模型在RMSE、MAE、MAPE三个评价指标的表现均为最优,展示了优化算法和神经网络在金融市场预测中的潜力,也为实际金融投资提供了一定的参考性。Accurate stock price prediction is of great significance for investor decision-making and financial institution risk management in the stock market. With the rapid development of computer technology, more and more financial scholars are paying attention to neural networks and using network technology to establish a stock closing price trend prediction model, which has achieved significant results. This article explores stock price prediction based on the Gated Recurrent Unit model selected in Matlab. Considering that the selection of neural network parameters has a significant impact on predictive performance, an improved hippopotamus optimization algorithm was used to optimize the GRU model parameters and construct an IHO-GRU model to predict the closing price of CITIC Bank. The experimental results showed that the IHO-GRU model performed the best in the RMSE, MAE, and MAPE evaluation indicators, demonstrating the potential of optimization algorithms and neural networks in financial market prediction, and providing some reference for actual financial investment.
