我国股市预测中ARIMA-NN混合模型与GARCH族模型的比较研究
发布时间：2018-11-12 16:19
【摘要】：股市是资本市场的重要组成部分,它是宏观经济最灵敏的晴雨表,经济的波动往往最先体现在股价涨跌行情中。此外,它在增加资本流动性,协调社会资源配置,提供投资渠道等方面也有举足轻重的作用。经过二十余年的迅速发展,我国股市已具有相当规模,截止2012年12月底,我国沪深两市市值合计达36973.7亿美元①,在世界各大证券交易所市值排名中占居第三位。 尽管我国股市在不断完善,相关法律法规也逐步健全,但仍存在一些特有的问题,如股票流通限制、政府导向性强、投机性高、投资者非理性等问题,因此经常表现出比世界成熟资本市场更大的波动性。而股票价格的非正常波动不仅会给投资者带来不必要损失,同时也会危害证券市场的健康发展以及我国整个经济体系的稳定性。由此可见,对我国股票价格指数进行深入研究具有重要意义。 由于金融时间序列往往同时具有线性和非线性特点,因此本文提出了将差分自回归移动平均(ARMA)模型与神经网络(NN)相结合的混合模型,同时本文分别运用ARIMA模型、NN模型、ARIMA-GARCH族模型和ARIMA-NN混合模型,对我国1996年12月16日至2009年12月31日的上证综指和深证成指收盘价格数据建立时间序列预测模型,并用2010年1月4日至2012年12月31日数据用于验证所建模型的准确性,通过比较预测结果,得出混合模型较其他模型更能提高我国股票价格预测准确率的相关结论。同时根据ARIMA-GARCH族模型结果,分析得出我国股票市场的波动性特征,并据此提出我国股市未来发展的相关建议。
[Abstract]:The stock market is an important part of the capital market, it is the most sensitive barometer of the macro economy, the economic fluctuations are often reflected in the stock price ups and downs. In addition, it also plays an important role in increasing capital mobility, coordinating social resource allocation and providing investment channels. After more than 20 years of rapid development, China's stock market has a considerable scale. By the end of December, 2012, the market value of Shanghai and Shenzhen stock markets amounted to 3.69737 trillion US dollars, ranking third among the world's major stock exchanges. Although China's stock market is constantly improving and relevant laws and regulations are gradually improving, there are still some unique problems, such as restrictions on the circulation of stocks, strong government guidance, high speculation, irrational investors, and so on. As a result, it often shows greater volatility than the world's mature capital markets. The abnormal fluctuation of stock price will not only bring unnecessary losses to investors, but also endanger the healthy development of the securities market and the stability of the whole economic system of our country. Thus, it is of great significance to study the stock price index in China. Because financial time series are always linear and nonlinear at the same time, a hybrid model combining differential autoregressive moving average (ARMA) model with neural network (NN) is proposed, and the ARIMA model is used in this paper. NN model, ARIMA-GARCH family model and ARIMA-NN mixed model are used to predict the closing price of Shanghai Composite Index and Shenzhen Composite Index from December 16, 1996 to December 31, 2009. The data from January 4, 2010 to December 31, 2012 are used to verify the accuracy of the proposed model. By comparing the prediction results, it is concluded that the hybrid model can improve the accuracy of stock price prediction in China better than other models. At the same time, according to the results of ARIMA-GARCH family model, the characteristics of volatility of Chinese stock market are analyzed, and the relevant suggestions for the future development of Chinese stock market are put forward.
【学位授予单位】：天津财经大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F832.51;F224
本文编号：2327574
[Abstract]:The stock market is an important part of the capital market, it is the most sensitive barometer of the macro economy, the economic fluctuations are often reflected in the stock price ups and downs. In addition, it also plays an important role in increasing capital mobility, coordinating social resource allocation and providing investment channels. After more than 20 years of rapid development, China's stock market has a considerable scale. By the end of December, 2012, the market value of Shanghai and Shenzhen stock markets amounted to 3.69737 trillion US dollars, ranking third among the world's major stock exchanges. Although China's stock market is constantly improving and relevant laws and regulations are gradually improving, there are still some unique problems, such as restrictions on the circulation of stocks, strong government guidance, high speculation, irrational investors, and so on. As a result, it often shows greater volatility than the world's mature capital markets. The abnormal fluctuation of stock price will not only bring unnecessary losses to investors, but also endanger the healthy development of the securities market and the stability of the whole economic system of our country. Thus, it is of great significance to study the stock price index in China. Because financial time series are always linear and nonlinear at the same time, a hybrid model combining differential autoregressive moving average (ARMA) model with neural network (NN) is proposed, and the ARIMA model is used in this paper. NN model, ARIMA-GARCH family model and ARIMA-NN mixed model are used to predict the closing price of Shanghai Composite Index and Shenzhen Composite Index from December 16, 1996 to December 31, 2009. The data from January 4, 2010 to December 31, 2012 are used to verify the accuracy of the proposed model. By comparing the prediction results, it is concluded that the hybrid model can improve the accuracy of stock price prediction in China better than other models. At the same time, according to the results of ARIMA-GARCH family model, the characteristics of volatility of Chinese stock market are analyzed, and the relevant suggestions for the future development of Chinese stock market are put forward.
【学位授予单位】：天津财经大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F832.51;F224
【参考文献】
相关期刊论文 前10条
1 梁艳春,聂义勇;从科学研究方法论看人工神经网络研究的发展[J];吉林大学学报(信息科学版);2002年01期
2 赵成柏;毛春梅;;基于ARIMA和BP神经网络组合模型的我国碳排放强度预测[J];长江流域资源与环境;2012年06期
3 邹辉文;黄明星;;基于股价异常波动的中国股市监管效率实证分析[J];财经研究;2010年01期
4 刘煜辉,贺菊煌,沈可挺;涨跌停板制度下中国A、B股市场波动的动态变化[J];管理科学;2003年05期
5 谢百三;;中国股市为何没有与实体经济同步发展[J];价格理论与实践;2012年05期
6 陈其安;张媛;刘星;;宏观经济环境、政府调控政策与股票市场波动性——来自中国股票市场的经验证据[J];经济学家;2010年02期
7 庞介民,王庆仁;新兴加转轨条件下中国证券公司的风险成因及监控[J];经济研究;2003年12期
8 赵俊强;廖士光;李湛;;中国上市公司股权分置改革中的利益分配研究[J];经济研究;2006年11期
9 刘华春;;基于Elman神经网络的股市决策模型[J];计算机应用;2009年S1期
10 蔡红;陈荣耀;;基于PCA-BP神经网络的股票价格预测研究[J];计算机仿真;2011年03期
，本文编号：2327574
本文链接：https://www.wllwen.com/jingjilunwen/zbyz/2327574.html
