基于时间序列分析的沪深股票价格预测 随着中国的市场经济的快速发展,股票市场也从三十年前的起步阶段，进入了快速发展的新阶段,除了实体经济,现在居民越来越青睐资本市场的金融产品,其中最受人们关注的金融资产就是股票,股票在人们的日常生活当中可以帮助人们创造财富，目前来看，投资权益类的资产已经成为人们获取报酬的有效方式；对于每一个企业而言，企业股价的波动可以反应出这个企业在过去一段时间的利润和收益以及公司经营风险情况。然而，股票价格的波动是一个十分复杂的类似于布朗运动的随机游走动态过程，通常情况下，我们不能对股价进行线性回归，因为从长期来看，股价不会呈现出线性趋势而是随机的波动。因此这要求我们寻找到新的方法研究股价的变化。我们发现时间序列模型能够有效并且较为精确地拟合曲线,因为它通过收集股票在过去一段时间的数据，来进行拟合分析。本文通过各种渠道收集了沪深三百指数从2018年1月至2020年12月的每日股票收盘价信息,利用时间序列模型对股票价格的变动进行分析建模预测。使用差分的方法对股票价格数据进行平稳化处理，通过单位根检验等方法对股票价格差分过后进行的平稳性分析,并用ARIMA模型进行拟合，研究发现，有的模型系数显著为0，因此要改变模型为疏系数模型进一步分析预测。从而对后期的股票价格变化趋势进行了更加准确的预测,之后对于预测结果进行分析发现沪深三百指数的股票价格受滞后第一期和第二期股票价格因素的影响较为有限,更多地受到其滞后的第三期和第四期的股票价格对当期的股票价格的影响；从长期来看，股价稳定，但与市场情况存在差异，所以我们只进行短期预测 关键词：时间序列分析；ARIMA；差分；单位根检验；疏系数模型 ABSTRACT With the fast and stable development of China's market economy, our stock market from the beginning of thirty years ago, has entered a new stage of rapid development, in addition to the real economy, now more and more favor of capital market financial products, the most attention to financial products is a stock, stock in People's Daily life can help people to create benefits. For listed companies, the fluctuation of stock price can reflect the profits and earnings of the company in the past period of time as well as the operating risks of the company. At the same time, investing in equity assets has become an effective way for people to get paid. It is often said that the stock market is the barometer of economic development, which indicates that the development of the stock market can indicate the performance of the economy. The development of stock is more and more closely related to the development of economy. The fluctuation of stock price can reflect the implementation of national economic policies and the living conditions of residents comprehensively. The fluctuation of stock price is a very complex dynamic process similar to the Brownian motion of random walk. In general, we cannot carry out linear regression for stock price, because in the long run, stock price will not show a linear trend but a random fluctuation. Therefore, this requires us to find new ways to study the changes in stock prices. Time series model can effectively fit the curve, because it studies the data of the stock in the past period of time to carry out the fitting analysis. This paper collected the daily stock closing price information of the CSI 300 Index from January 2018 to December 2020 through various channels, and used the time series model to analyze and model the changes of stock prices. Using the difference method to stabilize the stock price data, through the unit root test and other methods of the stock price difference after the stability analysis, and using the ARIMA model for fitting, the study found that some model coefficient is significantly 0, so we need to change the model to the model of the density of further analysis and prediction. Thus, on the stock price change trend in the late more accurate predictions, for predicting the result after analysis found that the CSI 300 index of stock prices by lag phase I and II influence factors of the stock price is limited, but its phase lag of the third and fourth stock price has a more powerful impact on the stock price in the future; In the long term, stock prices are stable, but they differ from the market, so we only make short-term forecasts Keywords: Time series analysis; difference; unit root test; sparse coefficient model 目录 一、引言 1 （一）研究背景 1 （二）研究目的及意义 1 二、时间序列分析的基本模型 2 （一）平稳时间序列所用到的模型 2 1、自回归（AR）模型 2 2、移动平均（MA）模型 2 3、移动平均自回归（ARMA）模型 2 4、三种类型的判别 2 5、ARMA模型的建模步骤 3 （二）非平稳时间序列所用到的模型 3 1、非平稳序列确定性因素分析 3 2、非平稳序列随机分析 3 3、ARIMA模型的结构 3 4、ARIMA模型建模步骤 4 三、模型建立 4 （一）获取数据并导入 4 （二）观察时序图序列是否平稳 6 （三）平稳性分析 7 （四）单位根检验 8 （五）白噪声检验 8 （六）拟合模型 8 （七）残差白噪声检验 9 （八）参数显著性检验 9 （九）疏系数模型建立 10 四、模型预测分析 11 五、总结与展望 12
*若需了解更多与协助请咨询↓→[电脑QQ][手机QQ]【数据协助】
