%0 Journal Article
%T 基于误差修正和CEEMDAN-IGWO-ELM的股票价格预测建模<br>Stock Price Prediction Modeling Based on Error Correction and CEEMDAN-IGWO-ELM
%A 谢汇钦
%A 颜七笙
%J Advances in Applied Mathematics
%P 2256-2273
%@ 2324-8009
%D 2024
%I Hans Publishing
%R 10.12677/aam.2024.135214
%X 针对股票价格非平稳、非线性等特性引发预测精度低的问题，引入Halton Sequence搜索算法、莱维飞行与等级制度策略对灰狼优化算法(GWO)进行改进，提出一种基于误差修正和CEEMDAN-IGWO-ELM股票价格预测模型。首先将股票交易数据进行归一化处理，作为极限学习机(ELM)的输入对股票价格进行预测得到初始预测结果，进而得到误差序列。然后利用PE自适应地确定自适应噪声完备集合经验模态分解(CEEMDAN)的参数，对误差序列进行分解，利用IGWO算法优化ELM模型可调参数对每个子序列建模预测，叠加各子序列预测结果对初始预测序列进行误差修正，得到最终股票预测值。仿真实验与Diebold-Mariano检验结果表明，与其他预测模型相比，所建立模型具有更高的预测精度和优越性。<br />
Aiming at the problem of low prediction accuracy caused by non-static and non-linear characteristics of stock price, the grey wolf optimization algorithm (GWO) is improved by introducing Halton Sequence search algorithm, Levy flight and hierarchy strategy. A stock price prediction model based on error correction and CEEMDAN-IGWO-ELM was proposed. Firstly, the stock transaction data was normalized as input for extreme learning machine (ELM) to predict the stock price to obtain the initial prediction result and obtain an error sequence. Then, PE is used to adaptively determine the parameters of complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and decompose the error sequence. The IGWO algorithm is used to optimize the adjustable parameters of ELM model and predict each subsequence. After stacking the prediction results of each, the Error subsequences obtain the final stock prediction result. The experimental results and Diebold-Mariano test show that compared with other prediction models, this model has better prediction accuracy and superiority.
%K 自适应噪声完备集合经验模态分解，灰狼优化算法，极限学习机，误差修正，股票价格预测<br>Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN)
%K Grey Worf Optimization (GWO)
%K Extreme Learning Machine (ELM)
%K Error Correction
%K Stock Price Prediction
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=88412