| 中文题名： | ARIMA-GARCH模型在交易型开放式指数基金中的应用 |
| 姓名： | |
| 保密级别： | 公开 |
| 论文语种： | 中文 |
| 学科代码： | 025200 |
| 学科专业： | |
| 学生类型： | 硕士 |
| 学位： | 应用统计硕士 |
| 学位类型： | |
| 学位年度： | 2020 |
| 校区： | |
| 学院： | |
| 第一导师姓名： | |
| 第一导师单位： | |
| 提交日期： | 2020-06-23 |
| 答辩日期： | 2020-06-23 |
| 外文题名： | APPLICATION OF ARIMA-GARCH MODEL IN TRADING OPEN INDEX FUND |
| 中文关键词： | |
| 中文摘要： |
近些年来，交易型开放式指数基金（ETF）在证券市场中扮演着越来越重要的作用。ETF 如此重要的原因就是其相对于其他证券形式有很多优势，有规避某些市场风险，收益稳健等诸多优势。因为 ETF 所受的高关注度及在我国发展过程中的地位，本文选取 ETF 来进行研究，由于 ETF 单位净值数据通过多次比较研究，内在规律发现，使用 ARIMA 模型来分析能够取得较效果。所以本文通过 ARIMA 模型来研究 ETF 单位净值的变化趋势，及模型适应性的内在规律。为了研究交易型开放式指数基金单位净值内在规律，本文选取三种具有代表性的 ETF 来进行研究，分别为创业板 ETF，300 ETF，纳指 ETF。本文最终发现 ARIMA 模型研究 ETF 取得较好效果，但运用之后发现了残差的集聚性，因而引用 GARCH 模型来解决。 针对创业板 ETF 单位净值数据，在进行严格的平稳性检验之后，对于存在的不平稳性，异方差等问题，先进行平稳化处理，进而使用平稳模型进行拟合，最终确定 ARIMA(1,1,10)并估计其系数，并进行了模型与系数的显著性，拟合之后，发现残差集聚性，并最终发现 GARCH(3,3)模型来解决相关性问题较好，并提高了模型的预测准确度。300 ETF 的分析研究与创业板 ETF 的研究方法类似。纳指 ETF 运用 AIC 准则判断 ARIMA 模型系数出现了模型系数不显著问题，因而结合 BIC 等准则最终选择了 ARIMA (1, 1, 8) 来对纳指 ETF 进行分析，残差序列使用 GARCH (2, 2) 来刻画，最终提高模型预测准确度。 根据本文研究的三种具有代表性的 ETF，本文提出一种投资策略，“双轨机制”的投资策略，利用 ARIMA 模型对 ETF 单位净值拟合之后的区间估计，模拟当前模型的准确度，预测未来 ETF 单位净值的变化趋势，进而判断此 ETF 的投资与否。 |
| 外文摘要： |
In recent years, trading open index funds (ETF) have played an increasingly important role in the securities market. The reason is that it has many advantages over other forms of securities, such as the avoidance of certain market risks and stable returns. Because of the high attention received by ETF and their status in China's development process, this article selects ETF for research. This article uses the ARIMA model to study the trend of ETF unit net value and the internal laws of model adaptability. In order to study the inherent law of the unit net value of trading open-ended index funds, this paper selects three representative trading open-end index funds for research, namely, the Growth Enterprise Market ETF, 300 ETF, and Nasdaq ETF. This article mainly uses the time series model to study. After many simulation analyses, this article finally uses the ARIMA model to study, but after using it, the autocorrelation of the residuals is found, so the GARCH model is used to solve it. For the ETF unit net value data, after conducting a rigorous stationarity test, for the problems of instability, heteroscedasticity,, first carry out a smoothing process, and then use a stationary model to fit, and finally determine ARIMA(1,1,10) and estimate the coefficients, and the significance of the model and the coefficients. After fitting, the multi-step correlation of the residuals is found, and the GARCH (3, 3) model is finally selected to solve the correlation problem and improve. The prediction accuracy of the model. The analysis and research of 300 ETF is similar to that of ETF. The Nasdaq ETF uses the AIC criterion to judge the ARIMA model coefficients. The model coefficients are not significant, so in combination with the BIC and other criteria, ARIMA (1, 1, 8) was finally selected to analyze the Nasdaq ETF. The residual sequence uses GARCH (2, 2) Come to describe, and ultimately improve the accuracy of model prediction. According to the three representative ETF studied in this paper, this paper proposes an investment strategy, the "dual-track mechanism" investment strategy, using the ARIMA model to estimate the interval after the ETF unit net value is fitted, to simulate the accuracy of the current model and predict the future The change trend of ETF unit net value, and then judge whether the ETF is invested or not.
﹀
|
| 馆藏号： | 硕025200/20034 |
| 开放日期： | 2021-06-23 |
谷歌
