随着人们在股票投资领域的不断深入，当前对于股票价格变化的趋势分析与预测研究亦在不断深化。由于影响股票价格变化的因素众多且极具不确定性，从而导致股票价格数据具有高噪音、非平稳、高波动等明显特征，进一步增加了分析与预测研究的困难性。基于这一背景，本文在总结前人已有研究理论与成果的基础上，构建基于数据特征提取及自适应参数调整下的KPCA-IWOA-LSTM模型，实现了股票价格变化的趋势分析与预测研究，具体工作开展如下：

首先在不同市场、不同属性标的下选取研究对象、确定数据指标、采集研究数据，并利用核主成分分析（KPCA）完成数据降维与特征提取。其次在自适应参数调整的思路下，利用改进的鲸鱼算法（IWOA）对长短期记忆神经网络进行参数的调整与优化，以寻找最优参数解，以此提升模型的运行效果。再次，将其所得到的最优参数组合传入已搭建的LSTM模型中，完成在LSTM神经网络下的模型运算。最后，通过实验设置的模型评价指标对各实验模型完成综合比对与分析，得出实验结论。

通过不同研究对象下不同模型的输出结果比较验证，本文所构建的KPCA-IWOA-LSTM模型的运行效果相较于RNN模型、LSTM模型、ACO-LSTM模型以及WOA-LSTM模型均有不同程度的提升。

实验结果表明，该模型使得股票数据特征与神经网络拓扑结构之间相互匹配，提高了整体预测能力。这也进一步说明本文的算法设计思路对股票价格变化的趋势分析与研究预测具有一定的借鉴价值。

With the continuous deepening of people's investment in the field of stock investment, the current trend analysis and forecasting research on stock price changes is also deepening. Due to the many factors affecting the stock price changes and the great uncertainty, which leads to the stock price data with high noise, non-stationary, high volatility and other obvious characteristics, which further increases the difficulty of analysis and prediction research. Based on this background, this paper summarises the existing research theories and results of previous researchers, and constructs the KPCA-IWOA-LSTM model based on data feature extraction and adaptive parameter adjustment, to achieve the trend analysis and prediction research of stock price changes, and carries out the specific work as follows:

Firstly, the research object is selected under different markets and different attribute targets, data indicators are determined, research data are collected, and data dimensionality reduction and feature extraction are completed by using Kernel Principal Component Analysis (KPCA). Secondly, under the idea of adaptive parameter adjustment, the Improved Whale Algorithm (IWOA) is used to adjust and optimise the parameters of the long and short-term memory neural network to find the optimal parameter solution, so as to improve the operating effect of the model. Again, the optimal parameter combinations obtained from it are passed into the constructed LSTM model to complete the model operation under the LSTM neural network. Finally, the experimental models are compared and analysed by the model evaluation indexes set up in the experiment, and conclusions are drawn.

Through the comparison of the output results of different models under different research objects, it is verified that the operation effect of the KPCA-IWOA-LSTM model constructed in this paper is improved to different degrees compared with the RNN model, LSTM model, ACO-LSTM model and WOA-LSTM model.

The experimental results show that the model makes the stock data features and the neural network topology match each other, which improves the overall prediction ability. It also further shows that the algorithm design ideas in this paper are valuable for trend analysis and research prediction of stock price changes.

[1] 钱津.论中国股票市场的高质量发展[J]. 经济与管理评论,2021,37(06):5-14. [2]

[2] 田永秀.中国近代股票市场研究[M].北京:人民出版社,2015.10-15.

[3] 毛越.投资者情绪与股市流动性之间的关系分析[J].现代商业,2022(10):3

[4] 王唯贤,陈利军.股票价格预测的建模与仿真研究[J].计算机仿真,2012(1):4.

[5] 何卓键.浅谈人工智能与机器学习[J].电子世界,2018,04No.538):185-186.

[6] 牛润苗.基于行为金融学视角对股市投资行为探析[J].经济与管理科学·金融,2019 (42):168- 173.

[7] 孟辰星.股票投资者风险偏好研究[J].经济体制改革,2011(1):6.

[8] 邢伟琛.大数据环境下的股票预测探究[J].中国商论,2020(3):2

[9] White. Economic Prediction Using Neural Networks: the Case of IBM Daily Stock Returns[C].International Conference on Neural Networks. IEEE, 1988.

[10]KOLARIK T,RUDORFER G.Time series forecasting using neural networks [J].Aem Sigapl Apl Quote Quad ,1994 ,25 (1)86-94.

[11]Peter G, Zhang. Time series forecasting using a hybrid ARIMA and neural network model[J]. Neurocomputing, 2003, 50(1): 159-175.

[12]Nur Ozkan-Gunay E, Ozkan M. Prediction of bank failures in emerging financial markets: an ANN approach[J]. The Journal of Risk Finance, 2007, 8(5): 465-480.

[13]Dounias G. A Hybrid Intelligent System for Financial Tim e-series Forecasting[J] International Journal of Engineering Intelligent Systems for Electrica Engineering and Comm unications, 2008, 16(4): 193-213.

[14]Adhikari R, Agrawal R K. A combination of artificial neural network and random walk models for financial time series forecasting[J]. Neural Computing and Applications, 2014, 24: 1441-1449.

[15]Zahed J, Rounaghi M M. Application of Artificial Neural Netw ork M odels and Principal Component Analysis Method in Predicting Stock Prices onTehran Stock Exchange[J]. Physica, A. Statistical Mechanics and its Applications, 2015, 438:178-187.

[16]Jahan I, Sajal S. Stock price prediction using Recurrent Neural Network (RNN) algorithm on time-series data[C]2018 Midwest Instruction and Computing Symposium. Duluth, Minnesota, USA: MSRP, 2018.

[17]Nelson D M Q, Pereira A C M, Oliveira R A D. Stock market's price movement prediction with LSTM neural networks[C]2017 International Joint Conference on Neural Networks (IJCNN). IEEE, 2017.

[18]Taewook K, Young K H. Forecasting Stock Prices with a Feature Fusion LSTM-CNN Model Using Different Representations of the Same D ata[J]. Pubmed, 2019, 14(2):1-23.

[19]孙瑞奇.基于 LSTM 神经网络的美股股指价格趋势预测模型的研究[D].北京:首都 经济贸易大学,2016.

[20]孙存浩,胡兵,邹雨轩.指数趋势预测的BP-LSTM 模型[J].四川大学学报(自然科学版),2020,57(01):27-31.

[21]林楠.基于 BP 神经网络和 GARCH 模型的中国银行股票价格预测实证分析[D].兰州:兰州大学,2014.

[22]吴立扬,周波.基于人工神经网络的股价指数预测[J].知识经济,2009 (1):62-63.

[23]黄婷婷,余磊. SDAE-LSTM 模型在金融时间序列预测中的应用[J]. 计算机工程与应用, 2019, 55(01): 142-148.

[24]刘玉敏, 李洋, 赵哲耘. 基于特征选择的 RF-LSTM 模型成分股价格趋势预测[J]. 统计与决策, 2021, 37(01):157-160.

[25]Marco Dorigo,Luca Maria Gambardella. Ant colonies for the travelling salesman problem[J]. BioSystems,1997,43(2).

[26]Vittorio Maniezzo,Antonella Carbonaro. An ANTS heuristic for the frequency assignment problem[J]. Future Generation Computer Systems,2000,16(8).

[27]Rong-Long WANG,Xiao-Fan ZHOU,Kozo OKAZAKI. Ant Colony Optimization with Genetic Operation and Its Application to Traveling Salesman Problem[J]. IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences,2009,E92.A(5).

[28]Boxin Guan,Yuhai Zhao,Yuan Li. An improved ant colony optimization with an automatic updating mechanism for constraint satisfaction problems[J]. Expert Systems With Applications,2021,164.

[29]Amorim Kelvin Santos,Pavani Gustavo Sousa. Ant Colony Optimization-based distributed multilayer routing and restoration in IP/MPLS over optical networks[J]. Computer Networks,2021,185.

[30]Zannou Abderrahim,Boulaalam Abdelhak,Nfaoui El Habib. Relevant node discovery and selection approach for the Internet of Things based on neural networks and ant colony optimization[J]. Pervasive and Mobile Computing,2021,70.

[31]V. Sangeetha,R. Krishankumar,K. S. Ravichandran,Samarjit Kar. Energy-efficient green ant colony optimization for path planning in dynamic 3D environments[J]. Soft Computing,2021

[32]章琦,庞小红,吴智铭.约束法蚁群算法在多目标VRP中的研究[J].计算机仿真,2007(03):262-265.

[33]黄茜.蚁群算法及其在 TSP 中的应用[D].重庆大学,2008.

[34]王书勤.车辆路径问题的蚁群算法研究[D].重庆大学,2008.

[35]任代蓉. 基于改进蚁群算法的传感器网络能量管理的研究[D]:电子科技大学,2008.

[36]Mirjalili S, Lewis A. The whale optimization algorithm[J]. Advances in engineering software, 2016, 95: 51-67.

[37]Aljarah I, Faris H, Mirjalili S.Optimizing connection weights inneural networks using the whale optimization algorithm[J]. Soft Computing, 2016: 1-15.

[38]Jangir P, Trivedi I N, Bhesdadiya R H, etc. Training Multi-Layer Perceptron in NeuralNetwork using Whale Optimization Algorithm[J]. Indian Journal of Science & Technology, 2016, 9(19): 28-36.

[39]Mostafa A, Hassanien A E, Houseni M, etc. Liver segmentation in MRI images based on whale optimization algorithm[J]. Multimedia Tools & Applications, 2017(6):1-24.

[40]Hassan G, Hassanien A E. Retinal fundus vasculature multilevel segmentation using whale optimization algorithm[J]. Signal Image & VideoProcessing, 2017(1): 1-8.

[41]Khaled ben oualid Medani, Samir Sayah, Abdelghani Bekrar. Whale optimization algorithm based optimal reactive power dispatch: a case study of the Algerian power system[J]. Electric Power Systems Research, 2018, 163: 696-705.

[42]Javid Ghahremani-Nahr, Ramez Kian, Ehsan Sabet. A robust fuzzy mathematical programming model for the closed-loop supply chain network designand a whale optimization solution algorithm[J]. Expert Systems With Application, 2019, 116:454-471.

[43]Dalia Yousri, Dalia Allam, M.B. Eteiba. Chaotic whale optimization variants for parameters estimation of the chaotic behavior in permanent magnet synchronous motor[J]. Applied Soft Computing, 2019, 74: 479-503.

[44]Yun Zhang, Yongguo Liu, Jieting Li, etc. WOCDA: A whale optimizationbased community detection algorithm[J]. Physica A, 2020, 539: 122937.

[45]崔东文.鲸鱼优化算法及其在水库优化调度中的应用研究[C]云南省水利学会.云南省水利学会 2018 年度学术交流会论文集.[出版详],2018:821-829.

[46]刘竹松,李生.正余混沌双弦鲸鱼优化算法[J].计算机工程与应用,2017,55(3):65-72.

[47]许瑜飞,钱锋,杨明磊等.改进鲸鱼优化算法及其在渣油加氢参数优化的应用[J]. 化工学报, 2017, 30(1): 20-25.

[48]巩世兵, 沈海斌. 仿生策略优化的鲸鱼算法研究[J]. 传感器与微系统, 2017, 36(12): 10-12

[49]MIRJALILI S,LEWIS A. The w hale optimization algorithm[J].Advances in engineering software,2016,95:51-67.

[50]刘磊, 白克强, 但志宏, 张松, 刘知贵. 一种全局搜索策略的鲸鱼优化算法[J]. 小型微型计算机系统, 2020, 41(09): 1820-1825.

[51]钱琦淼,张朵,王映朦,等.机器学习在股价预测上的应用[J].中国市场,2022(21):4.

[52]彭燕,刘宇红,张荣芬.基于 LSTM 的股票价格预测建模与分析[J].计算机工程与应用,2019,55(11):209-212.

[53]唐小初.面向股票聚类的相似度和评价指标的研宄[D].深圳:深圳大学,2019.