基于HMM-SARIMA模型的选股策略——以沪深300医药行业成分股为例  

Stock Selection Strategy Based on HMM-SARIMA Model——A Case Study of Shanghai and Shenzhen 300 Pharmaceutical Stocks

作　　者：张骅月 牛艺潼 Zhang Huayue;Niu Yitong

机构地区：[1]南开大学金融学院

出　　处：《工程经济》2024年第5期13-30,共18页ENGINEERING ECONOMY

摘　　要：股票市场变化莫测,股市价格波动比较大,建立科学的模型来分析市场以及股票状态,针对不同的投资策略选择合适的股票组合是有必要的。本文提出了一种基于隐马尔可夫(HMM)模型和时间序列模型的股票状态预测方法,该方法可以有效地分析股票市场的状态,再针对不同的市场状态对股市价格分别进行预测,两种模型相结合的方法能够有效避免市场波动对单一模型预测结果的影响,最后根据模型的预测结果并针对不同的投资策略为投资者选择合适的股票,从而获得更大的收益。以与样本期不交叉的沪深300指数分行业成分股为样本,综合考虑收益和风险因素,为股票池中的每只股票进行打分,选出最优的投资组合,本文的模型在股市行情不稳定的回溯期内取得了很好的收益效果,这说明将HMM模型与时间序列结合的模型具有一定的稳健性。The stock market is unpredictable,characterized by significant price fluctuations.Establishing scientific models to analyze market and stock statuses and selecting appropriate stock combinations for different investment strategies is crucial.This paper proposes a stock status prediction method based on Hidden Markov Model(HMM)and time series models,which effectively analyzes the stock market's status.Subsequently,it predicts stock prices based on different market conditions.Combining these two models helps to mitigate the impact of market volatility on individual model predictions.Finally,based on the model's predictions and different investment strategies,suitable stocks are selected for investors to maximize returns.Using the industry component stocks of the Shanghai and Shenzhen 300 Index that do not overlap with the sample period as samples,considering both profitability and risk factors,each stock in the stock pool is scored to select the optimal investment portfolio.Our model achieves good returns during the volatile market conditions of the retrospective period,demonstrating the robustness of combining Hidden Markov Model and time series models.

关 键 词：HMM模型 时间序列 投资组合 股价预测 

分 类 号：F832.51[经济管理—金融学] F426.72