本文的研究基于对浦东金桥股票1799个交易日的时间序列数据进行分析,旨在利用ARIMA-LSTM组合模型对该股票的收盘价进行精确预测。通过网格调参确定了ARIMA模型的最佳阶数,以确保模型能够有效捕捉时间序列数据中的趋势和周期性变化。利...本文的研究基于对浦东金桥股票1799个交易日的时间序列数据进行分析,旨在利用ARIMA-LSTM组合模型对该股票的收盘价进行精确预测。通过网格调参确定了ARIMA模型的最佳阶数,以确保模型能够有效捕捉时间序列数据中的趋势和周期性变化。利用这一优化后的ARIMA模型对浦东金桥的收盘价进行预测,从中获得预测残差。而后将残差数据输入到LSTM (长短期记忆网络)模型中。LSTM作为一种适合处理序列数据的深度学习模型,能够更好地捕捉数据中的长期依赖关系和非线性动态。通过结合ARIMA模型的残差和LSTM模型的预测能力,构建了一个ARIMA-LSTM组合模型,进一步提升了对浦东金桥收盘价未来走势的预测准确性和稳定性。This study is based on the analysis of 1799 trading days’ time series data of Pudong Jinqiao stock, aiming to use the ARIMA-LSTM combination model to accurately predict the closing price of the stock. The optimal order of the ARIMA model was determined through grid tuning to ensure that the model can effectively capture trends and periodic changes in time series data. Use this optimized ARIMA model to predict the closing price of Pudong Jinqiao and obtain the prediction residual. Then, the residual data will be inputted into the LSTM (Long Short Term Memory Network) model. LSTM, as a deep learning model suitable for processing sequential data, can better capture long-term dependencies and nonlinear dynamics in the data. By combining the residual of the ARIMA model with the predictive ability of the LSTM model, an ARIMA-LSTM combination model was constructed to further improve the accuracy and stability of predicting the future trend of the closing price of Pudong Jinqiao.展开更多
在经济快速发展和金融市场波动的背景下,股价预测对于投资者、企业管理者以及金融机构来说至关重要。这不仅有助于他们掌握未来的风险,还为投资决策和监管提供了重要依据。单一预测模型在股价预测中很难同时捕获到数据序列中的线性和非...在经济快速发展和金融市场波动的背景下,股价预测对于投资者、企业管理者以及金融机构来说至关重要。这不仅有助于他们掌握未来的风险,还为投资决策和监管提供了重要依据。单一预测模型在股价预测中很难同时捕获到数据序列中的线性和非线性特征,因此预测效果不理想。针对该问题提出了一种基于ARIMA模型与LSTM模型相结合的股价预测模型,综合考虑线性与非线性特征的股价预测。本文采用贵州茅台(600519) 2018年1月2日到2023年12月29日之间的每个交易日的日收盘价进行实验,实验结果表明,与单一的ARIMA模型和LSTM模型相比,ARIMA-LSTM组合模型在股价预测方面取得了较好的效果。In the context of rapid economic development and fluctuating financial markets, stock price prediction is crucial for investors, corporate managers, and financial institutions. It not only helps them assess future risks but also provides essential support for investment decisions and regulatory actions. A single predictive model often struggles to capture both linear and nonlinear features in stock price data, leading to suboptimal forecasting results. To address this issue, a hybrid stock price prediction model combining the ARIMA model and the LSTM model is proposed, which comprehensively considers both linear and nonlinear characteristics. This study uses the daily closing prices of Kweichow Moutai (600519) from January 2, 2018, to December 29, 2023, for experiments. The experimental results show that the ARIMA-LSTM hybrid model achieves better stock price prediction performance compared to using either the ARIMA or LSTM model alone.展开更多
文摘本文的研究基于对浦东金桥股票1799个交易日的时间序列数据进行分析,旨在利用ARIMA-LSTM组合模型对该股票的收盘价进行精确预测。通过网格调参确定了ARIMA模型的最佳阶数,以确保模型能够有效捕捉时间序列数据中的趋势和周期性变化。利用这一优化后的ARIMA模型对浦东金桥的收盘价进行预测,从中获得预测残差。而后将残差数据输入到LSTM (长短期记忆网络)模型中。LSTM作为一种适合处理序列数据的深度学习模型,能够更好地捕捉数据中的长期依赖关系和非线性动态。通过结合ARIMA模型的残差和LSTM模型的预测能力,构建了一个ARIMA-LSTM组合模型,进一步提升了对浦东金桥收盘价未来走势的预测准确性和稳定性。This study is based on the analysis of 1799 trading days’ time series data of Pudong Jinqiao stock, aiming to use the ARIMA-LSTM combination model to accurately predict the closing price of the stock. The optimal order of the ARIMA model was determined through grid tuning to ensure that the model can effectively capture trends and periodic changes in time series data. Use this optimized ARIMA model to predict the closing price of Pudong Jinqiao and obtain the prediction residual. Then, the residual data will be inputted into the LSTM (Long Short Term Memory Network) model. LSTM, as a deep learning model suitable for processing sequential data, can better capture long-term dependencies and nonlinear dynamics in the data. By combining the residual of the ARIMA model with the predictive ability of the LSTM model, an ARIMA-LSTM combination model was constructed to further improve the accuracy and stability of predicting the future trend of the closing price of Pudong Jinqiao.
文摘在经济快速发展和金融市场波动的背景下,股价预测对于投资者、企业管理者以及金融机构来说至关重要。这不仅有助于他们掌握未来的风险,还为投资决策和监管提供了重要依据。单一预测模型在股价预测中很难同时捕获到数据序列中的线性和非线性特征,因此预测效果不理想。针对该问题提出了一种基于ARIMA模型与LSTM模型相结合的股价预测模型,综合考虑线性与非线性特征的股价预测。本文采用贵州茅台(600519) 2018年1月2日到2023年12月29日之间的每个交易日的日收盘价进行实验,实验结果表明,与单一的ARIMA模型和LSTM模型相比,ARIMA-LSTM组合模型在股价预测方面取得了较好的效果。In the context of rapid economic development and fluctuating financial markets, stock price prediction is crucial for investors, corporate managers, and financial institutions. It not only helps them assess future risks but also provides essential support for investment decisions and regulatory actions. A single predictive model often struggles to capture both linear and nonlinear features in stock price data, leading to suboptimal forecasting results. To address this issue, a hybrid stock price prediction model combining the ARIMA model and the LSTM model is proposed, which comprehensively considers both linear and nonlinear characteristics. This study uses the daily closing prices of Kweichow Moutai (600519) from January 2, 2018, to December 29, 2023, for experiments. The experimental results show that the ARIMA-LSTM hybrid model achieves better stock price prediction performance compared to using either the ARIMA or LSTM model alone.
