金融业已成为我国经济的关键支柱，证券市场作为其重要组成部分，对经济发展有着深远影响。近年来，我国经济持续高质高效增长，政府加大证券市场支持力度，企业和个人纷纷参与证券贸易，推动了证券市场繁荣发展。股票作为证券市场的核心，其价格变化与多种因素密切相关，具有很大的不确定性和复杂性，在此背景下，股票价格预测问题也成了研究热点之一，采用有效的方法，实现股票价格的精准预测具有重要意义。针对股价预测研究在大型综合股指中存在的预测精度不高和拟合效果较差等问题，本文构建了新的RF-MITSO-LSSVR股价预测模型并对其进行了系统研究，具体如下。

首先，针对标准金枪鱼群优化算法（TSO）易早期收敛和陷入局部最优解的缺陷，提出了一种将佳点集种群初始化，莱维飞行改进螺旋觅食和自适应概率阈值三种改进策略相结合的多策略改进金枪鱼群优化算法（MITSO）。并与TSO算法、GA算法、PSO算法和GWO算法在基准测试函数上进行了性能对比，证明了MITSO算法具备更优秀的搜索机制和摆脱局部最优解的能力。

其次，选取包括股权风险溢价率等风险指标在内的16个股价相关指标作为预测模型输入变量，并采用随机森林（RF）对输入变量进行降维处理以消除指标间多重共线性的影响，提升模型训练效率。同时，针对输入变量中存在的非线性关系、多特征耦合等特点，选择了LSSVR作为基准模型，并且针对机器学习中人工设置超参数的局限性，采用本文提出的MITSO算法对最小二乘支持向量机回归（LSSVR）的正则化参数和核函数参数进行了优化，建立了基于RF-MITSO-LSSVR的股价预测模型。

最后，选择沪深300指数这一典型综合股指的历史数据进行了三组对比实验，对模型的准确性和有效性进行了检验，结果表明：相比于基准模型，RF-MITSO-LSSVR 模型具有更好的预测性能。同时，与TSO、GA、PAO、GWO算法相比，MITSO算法在LSSVR超参数的搜索上效果更好且对预测效果有一定提升。此外，与常用于股价预测的模型ARIMA和LSTM相比，本文构建的RF-MITSO-LSSVR 模型在综合型股指价格预测方面的精度与耗时上取得了平衡，具有一定的优势。

The financial industry has become a key pillar of China’s economy, with the stock market as an important component having a profound impact on economic development. In recent years, China’s economy has continued to grow with high quality and efficiency. The government has increased its support for the stock market, and both enterprises and individuals have actively participated in securities trading, promoting the prosperity and development of the stock market. As the core of the stock market, stock prices are closely related to various factors, characterized by great uncertainty and complexity. Against this backdrop, stock price prediction has also become one of the hot research topics. It is of great significance to adopt effective methods to achieve accurate prediction of stock prices. In response to the issues of low prediction accuracy and poor fitting effects in large comprehensive stock indices, this paper constructs a new RF-MITSO-LSSVR stock price prediction model and systematically studies it, as follows.

Firstly, in order to solve the shortcomings of the standard tuna swarm optimization algorithm (TSO), which is easy to converge early and fall into the local optimal solution, a multi-strategy improved tuna swarm optimization algorithm (MITSO) was proposed, which combined three improved strategies: population initialization of the best spot set, improved spiral foraging by Levy flight and adaptive probability threshold. The performance of the MITSO algorithm,the TSO algorithm,the GA algorithm, the PSO algorithm and the GWO algorithm is compared with the benchmark function, which proves that the MITSO algorithm has a better search mechanism and the ability to get rid of the local optimal solution.

Secondly, 16 stock price-related indicators, including equity risk premium rate and other risk indicators, were selected as the input variables of the prediction model, and random forest (RF) was used to reduce the dimensionality of the input variables to eliminate the influence of multicollinearity between indicators and improve the training efficiency of the model. At the same time, in view of the nonlinear relationship and multi-feature coupling in the input variables, LSSVR was selected as the benchmark model, and in view of the limitations of manual hyperparameters in machine learning, the regularization parameters and kernel function parameters of Least Squares Support Vector Machine Regression (LSSVR) were optimized by using the MITSO algorithm proposed in this paper, and a stock price prediction model based on RF-MITSO-LSSSVR was established.

Finally, historical data of the CSI 300 Index, a typical comprehensive stock index, are selected for three sets of comparative experiments to test the accuracy and effectiveness of the model. The results show that the RF-MITSO-LSSVR model has better predictive performance compared to the baseline model. Moreover, compared with TSO, GA, PAO, and GWO algorithms, MITSO algorithm is more efficient in searching for LSSVR hyperparameters and has improved the prediction effect to a certain extent. In addition, compared with the commonly used stock price prediction models ARIMA and LSTM, the RF-MITSO-LSSVR model constructed in this paper achieves a balance between accuracy and time consumption in the prediction of comprehensive stock index prices, demonstrating certain advantages.

[1]Shah D, Isah H, Zulkernine F. Stock market analysis: A review and taxonomy of prediction techniques[J]. International Journal of Financial Studies, 2019, 7(2): 26.

[2]Gu S, Kelly B, Xiu D. Empirical asset pricing via machine learning[J]. The Review of Financial Studies, 2020, 33(5): 2223-2273.

[3]卜林,孙丽玲,李政.地缘政治风险、经济政策不确定性与股票市场波动[J].南开经济研究,2020(05):185-205.DOI:10.14116/j.nkes.2020.05.010.

[4]李斌,龙真.中国股票市场可预测性研究：基于机器学习的视角[J].管理科学学报,2023,26(10):138-158.DOI:10.19920/j.cnki.jmsc.2023.10.008.

[5]Kumar D, Sarangi P K, Verma R. A systematic review of stock market prediction using machine learning and statistical techniques[J]. Materials Today: Proceedings, 2022, 49: 3187-3191.

[6]Lendasse A, de Bodt E, Wertz V, et al. Non-linear financial time series forecasting-Application to the Bel 20 stock market index[J]. European Journal of Economic and Social Systems, 2000, 14(1): 81-91.

[7]Kim K. Financial time series forecasting using support vector machines[J]. Neurocomputing, 2003, 55(1-2): 307-319.

[8]Sheta A F, Ahmed S E M, Faris H. A comparison between regression, artificial neural networks and support vector machines for predicting stock market index[J].Soft Computing, 2015, 7(8): 55-63.

[9]Suykens J A K, Vandewalle J. Least squares support vector machine classifiers[J]. Neural processing letters, 1999, 9: 293-300.

[10]Frohlich H, Chapelle O, Scholkopf B. Feature selection for support vector machines by means of genetic algorithm[C].Proceedings. 15th IEEE International Conference on Tools with Artificial Intelligence. IEEE, 2003: 142-148.

[11]Henrique, B. M., Sobreiro, V. A., & Kimura, H. Stock price prediction using support vector regression on daily and up to the minute prices[J].The Journal of Finance and Data Science,2018, 4(3), 183–201.

[12]Nava, N., Di Matteo, T., and Aste, T. (2018). Financial timeseries forecasting using empirical mode decomposition and support vector regression[J].Risks, 6(7):1-21.

[13]De Brabanter K, Karsmakers P, Ojeda F, et al. LS-SVMlab toolbox user's guide: version 1.7[M]. Katholieke Universiteit Leuven, 2010.

[14]Huang C H, Yang F H, Lee C P. The strategy of investment in the stock market using modified support vector regression model[J]. Scientia Iranica, 2018, 25(3): 1629-1640.

[15]Ou P, Wang H. Financial volatility forecasting by least square support vector machine based on GARCH, EGARCH and GJR models: evidence from ASEAN stock markets[J]. International Journal of Economics and Finance, 2010, 2(1): 51-64.

[16]冯宇旭,李裕梅.基于LSTM神经网络的沪深300指数预测模型研究[J].数学的实践与认识,2019,49(07):308-315.

[17]Nayak R K, Mishra D, Rath A K. A Naïve SVM-KNN based stock market trend reversal analysis for Indian benchmark indices[J]. Applied Soft Computing, 2015, 35: 670-680.

[18]Gowthul Alam M M, Baulkani S. Local and global characteristics-based kernel hybridization to increase optimal support vector machine performance for stock market prediction[J]. Knowledge and Information Systems, 2019, 60(2): 971-1000.

[19]Wang Y, Wang Y. Using social media mining technology to assist in price prediction of stock market[C].2016 IEEE International conference on big data analysis (ICBDA). IEEE, 2016: 1-4.

[20]Jia-Yen Huang,Jin-Hao Liu.Using social media mining technology to improve stock price forecast accuracy[J].104-116,2020,39(1):104-116.

[21]Huang C L, Tsai C Y. A hybrid SOFM-SVR with a filter-based feature selection for stock market forecasting[J]. Expert Systems with applications, 2009, 36(2): 1529-1539.

[22]王冰玉;刘勇军. 基于信噪比的KPCA-SVM-KNN算法的股价预测研究 [J]. 计算机与数字工程, 2022, 50 (04): 685-690.

[23]Puspitasari D A, Rustam Z. Application of SVM-KNN using SVR as feature selection on stock analysis for Indonesia stock exchange[C].AIP Conference Proceedings. AIP Publishing, 2018, 2023(1)：1-7

[24]张倩. 基于降维和机器学习模型的沪深300指数预测[D].江西：江西财经大学,2023.DOI:10.27175/d.cnki.gjxcu.2022.001769.

[25]Yang L, Shami A. On hyperparameter optimization of machine learning algorithms: Theory and practice[J]. Neurocomputing, 2020, 415: 295-316.

[26]Bischl B, Binder M, Lang M, et al. Hyperparameter optimization: Foundations, algorithms, best practices, and open challenges[J]. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 2023, 13(2),e1484:1-43.

[27]Hoque K E, Aljamaan H. Impact of hyperparameter tuning on machine learning models in stock price forecasting[J]. IEEE Access, 2021, 9: 163815-163830.

[28]Yun K K, Yoon S W, Won D. Prediction of stock price direction using a hybrid GA-XGBoost algorithm with a three-stage feature engineering process[J]. Expert Systems with Applications, 2021, 186: 115716.

[29]孙秋韵,刘金清,刘引等.基于改进GA参数优化的SVR股价预测模型[J].计算机系统应用,2015,24(09):29-34.

[30]Bazrkar M J, Hosseini S. Predict stock prices using supervised learning algorithms and particle swarm optimization algorithm[J]. Computational Economics, 2023, 62(1): 165-186.

[31]Hegazy O, Soliman O S, Salam M A. LSSVM-ABC algorithm for stock price prediction[J]. International Journal of Computer Trends and Technology (IJCTT) 2(2014):81-92.

[32]Hegazy O, Soliman O S, Salam M A. Comparative study between FPA, BA, MCS, ABC, and PSO algorithms in training and optimizing of LS-SVM for stock market prediction[J]. International Journal of Advanced Computer Research, 2015, 5(18): 35-45.

[33]陈晓姨.基于贝叶斯优化的CNN-LSTM股价预测模型[D].兰州：兰州大学,2023.DOI:10.27204/d.cnki.glzhu.2023.002202.

[34]Zhang Y, Yang S. Prediction on the highest price of the stock based on PSO-LSTM neural network[C].2019 3rd International Conference on Electronic Information Technology and Computer Engineering (EITCE). IEEE, 2019: 1565-1569.

[35]杨青沙. 基于粒子群优化算法和LSSVM模型的股价预测研究[D].厦门：厦门大学,2020.DOI:10.27424/d.cnki.gxmdu.2019.000130.

[36]Gong X L, Liu X H, Xiong X, et al. Forecasting stock volatility process using improved least square support vector machine approach[J]. Soft computing, 2019, 23: 11867-11881.

[37]Tian Z, Pei L. Stocks price prediction based on optimized echo state network by sparrow search algorithm[J]. International Journal of Dynamics and Control, 2024: 1-14.

[38]Swathi T, Kasiviswanath N, Rao A A. An optimal deep learning-based LSTM for stock price prediction using twitter sentiment analysis[J]. Applied Intelligence, 2022, 52(12): 13675-13688.

[39]Bazrkar M J, Hosseini S. Predict stock prices using supervised learning algorithms and particle swarm optimization algorithm[J]. Computational Economics, 2023, 62(1): 165-186.

[40]Li X, Khishe M, Qian L. Evolving deep gated recurrent unit using improved marine predator algorithm for profit prediction based on financial accounting information system[J]. Complex & Intelligent Systems, 2023: 1-17.

[41]Liu C, Gao F, Zhao Q, et al. The optimized gate recurrent unit based on improved evolutionary algorithm to predict stock market returns[J]. RAIRO-Operations Research, 2023, 57(2): 743-759.

[42]Bai C. AGA-GRU: An optimized GRU neural network model based on adaptive genetic algorithm[C].Journal of Physics: Conference Series. IOP Publishing, 2020,1651(1),012146

[43]Breiman L. Random forests[J]. Machine learning, 2001, 45: 5-32.

[44]Cortes C, Vapnik V. Support-vector networks[J]. Machine learning, 1995, 20: 273-297.

[45]Ariyo A A, Adewumi A O, Ayo C K. Stock price prediction using the ARIMA model[C].2014 UKSim-AMSS 16th international conference on computer modelling and simulation. IEEE, 2014: 106-112.

[46]Hochreiter S, Schmidhuber J. Long short-term memory[J]. Neural computation, 1997, 9(8): 1735-1780.

[47]Xie L, Han T, Zhou H, et al. Tuna swarm optimization: a novel swarm-based metaheuristic algorithm for global optimization[J]. Computational intelligence and Neuroscience, 2021, 2021: 1-22.

[48]Holland J H. Genetic algorithms[J]. Scientific american, 1992, 267(1): 66-73.

[49]Wang D, Tan D, Liu L. Particle swarm optimization algorithm: an overview[J]. Soft computing, 2018, 22: 387-408.

[50]Negi G, Kumar A, Pant S, et al. GWO: a review and applications[J]. International Journal of System Assurance Engineering and Management, 2021, 12: 1-8.

[51]GEZİCİ H. Improved Tuna Swarm Optimization Algorithm for Engineering Design Problems[J]. Journal of Materials and Mechatronics: A, 2023, 4(2): 424-445.

[52]华罗庚, 王元. 数论在近似分析中的应用 [ M ]. 北京科学出版社, 1978.

[53]李涵,李文敬.混合策略改进的金枪鱼群优化算法[J].广西科学,2023,30(01):208-218.DOI:10.13656/j.cnki.gxkx.20230308.022.

[54]吴继春,张斋武,杨永达,等.基于改进金枪鱼群算法的机械臂时间最优轨迹规划[J/OL].计算机集成制造系统,1-19[2024-04-03].https://doi.org/10.13196/j.cims.2023.0528.

[55]Tan M L, Li Y T, Ding D L, et al. An Improved JADE Hybridizing with Tuna Swarm Optimization for Numerical Optimization Problems[J]. Mathematical Problems in Engineering, 2022, 2022:1-17.

[56]Yao X, Liu Y, Lin G. Evolutionary programming made faster[J]. IEEE Transactions on Evolutionary computation, 1999, 3(2): 82-102.

[57]Hu P, Chen S, Huang H, et al. Improved alpha-guided grey wolf optimizer[J]. IEEE Access, 2018, 7: 5421-5437.

[58]Albadr M A, Tiun S, Ayob M, et al. Genetic algorithm based on natural selection theory for optimization problems[J]. Symmetry, 2020, 12(11),1758:1-31.

[59]Zhang X, Hu Y, Xie K, et al. A causal feature selection algorithm for stock prediction modeling[J]. Neurocomputing, 2014, 142: 48-59.

[60]Gündüz H, Çataltepe Z, Yaslan Y. Stock daily return prediction using expanded features and feature selection[J]. Turkish Journal of Electrical Engineering and Computer Sciences, 2017, 25(6): 4829-4840.

[61]Leippold M, Wang Q, Zhou W. Machine learning in the Chinese stock market[J]. Journal of Financial Economics, 2022, 145(2): 64-82.

[62]李程,张恒玮.经济政策不确定性、国债利率期限结构与股市收益率[J].金融市场研究,2024(03):98-111.DOI:10.20134/j.cnki.fmr.2024.03.005.

[63]姚发中.市盈率和收益率指标与股价预测[J].对外经贸实务,1994,(06):46-47.

[64]Hasan M A M, Nasser M, Ahmad S, et al. Feature selection for intrusion detection using random forest[J]. Journal of information security, 2016, 7(3): 129-140.

[65]Hu G X, Chen C, Shao Y, et al. Fama–French in China: size and value factors in Chinese stock returns[J]. International Review of Finance, 2019, 19(1): 3-44.

[66]Sirignano J, Cont R. Universal features of price formation in financial markets: perspectives from deep learning[M].Machine Learning and AI in Finance. Routledge, 2021: 5-15.

[67]Lu W, Li J, Wang J, et al. A CNN-BiLSTM-AM method for stock price prediction[J]. Neural Computing and Applications, 2021, 33(10): 4741-4753.

[68]Shen J, Shafiq M O. Short-term stock market price trend prediction using a comprehensive deep learning system[J]. Journal of big Data, 2020, 7: 1-33.