基于鲸鱼算法优化LSTM的股票价格预测模型  

Stock price forecast model based on whale algorithm optimizing LSTM

作　　者：李婧琦 LI Jingqi(Institute of statistics,Shandong Institute of Business and Technology,Yantai Shandong 264005,China)

机构地区：[1]山东工商学院统计学院,山东烟台264005

出　　处：《智能计算机与应用》2023年第2期35-40,共6页Intelligent Computer and Applications

摘　　要：长短期记忆神经网络(LSTM)因其长时记忆的可预测性,在金融领域脱颖而出。然而前期研究结果显示,该方法存在主观性决定关键参数,容易陷入局部最优,导致能力不佳的问题。基于上述问题,本文提出一种基于鲸鱼算法(WOA)优化长短期记忆网络(LSTM)的股票价格预测模型。该模型通过鲸鱼算法,对LSTM网络的重要参数进行寻优,使之降低人为因素的影响,提高模型预测的准确性。同时,针对股票数据之间的冗余性导致模型效率降低的问题,使用递归特征消除算法对数据进行特征选择,建立完善指标体系进行预测。实验以上证指数股票数据构建了WOA-LSTM模型,并对该模型的预测结果与单一LSTM、PSO-LSTM、SSA-BP模型进行比较分析。实验结果表明,本文所提模型对股票价格的预测明显优于其它模型。Long Short-Term Memory neural network stands out in finance for their long-term memory predictability.However,the previous research results show that the method has the subjective determination of key parameters,and it is easy to fall into the problem of poor performance due to local optimization.To solve above questions,this paper proposes a stock price prediction model based on the whale algorithm to optimize the Long Short-Term Memory network.The model uses the whale algorithm to optimize the important parameters of the LSTM network,which reduces the influence of human factors and improves the accuracy of model prediction.At the same time,in view of the problem that the redundancy between stock data leads to the reduction of model efficiency,recursive feature elimination algorithm is used to perform feature selection on the data,and a perfect indicator system is established for prediction.The WOA-LSTM model is constructed experimentally based on the stock data of the Shanghai Stock Exchange,and the prediction results of the model are compared and analyzed with the single LSTM,PSO-LSTM and SSA-BP models.Experimental results show that the proposed model is significantly better than other models in predicting stock prices.

关 键 词：鲸鱼优化 股票价格预测 长短期记忆网络 递归特征消除 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]