时间序列分析方法在黄金价格走势预测中的应用研究

发布时间：2018-01-04 04:24

本文关键词：时间序列分析方法在黄金价格走势预测中的应用研究　出处：《国防科学技术大学》2014年硕士论文　论文类型：学位论文

【摘要】：黄金是贵金属和硬通货。投资黄金可以保障投资者的资产不会被通货膨胀侵蚀的同时,也不易像股票投资和房地产投资那样会面对市场崩盘,因此深受人们所喜爱。针对黄金的历史价格数据采用数据挖掘,可以有效的预测黄金价格走势,指导投资者的投资决策,具有广泛的应用价值。本文针对黄金价格走势预测的应用需求,深入分析黄金价格的时间序列特征,总结出其除了包含常见的非线性,非平稳和动态等特征外,还具有高噪音和非正态等特点。介绍了国内外研究学者对黄金价格分析研究的现状,然后利用时间序列的相关原理和统计工具软件进行定量解释。本文对黄金价格进行预测,其创新性的应用主要表现在以下三个方面:1)基于灰色模型的原理,针对黄金价格走势高噪音、非正态的特点,提出了一种基于灰色模型的黄金价格走势预测方法,并将其应用到黄金价格的预测中,通过实际数据验证了该方法的有效性。2)基于差分自回归滑动平均模型的原理,针对黄金价格走势的内在特征,提出了一种基于差分自回归滑动平均模型的黄金价格走势预测方法,并将其应用到黄金价格的预测领域,通过实际数据验证了该方法的有效性。3)基于真实的黄金价格数据,对比分析了加权移动平均、指数平滑、灰色模型和ARIMA等模型对黄金价格走势预测的准确性和有效性,并归纳总结了这四种模型的优缺点和适用场景。文中主要采用基于灰色模型的黄金价格走势预测、基于ARIMA模型的黄金价格走势预测、移动平均模型以及指数平滑模型,并对模型的预测结果进行对比分析。通过对四种模型拟合的效果对比分析,基本上都可以描述黄金价格的走势,尤其求和自回归滑动平均模型拟合效果最佳。通过实际数据验证研究表明,该模型在短期预测方面有很强的实用价值。在此基础上以及其它核心技术的研究与突破,最后实现了整个系统的应用。
[Abstract]:Gold is a precious metal and hard currency investment. Investors can guarantee gold assets to be eroded by inflation at the same time, it is not easy as the stock and real estate investment that will face the market crash, so deeply loved by the people. According to the historical price data of gold by data mining, can effectively predict the price of gold trend, guide the investment decisions of investors, and has wide application value. Based on the application demand prediction of gold price movements, in-depth analysis of the characteristics of time series of gold price, summed up its non linear except contains common, non stationary and dynamic characteristics, but also has high noise and non normal characteristics. This paper introduces the status quo of the research researchers at home and abroad on the analysis of the price of gold, and then explain quantitatively using time series correlation principle and statistical software. In this paper, the price of gold in the pre The application of the innovative measure, mainly in the following three aspects: 1) principle based on grey model, the gold price trend of high noise characteristics of non normal, the paper proposed a prediction method of grey model based on the gold price trend, and its application to the prediction of gold price, through the actual data to verify the effectiveness of the method was.2) principle of ARIMA model based on the inherent characteristics of the gold price movements, we propose a autoregressive moving average model of the gold price trend forecast method based on difference, and its application to the prediction of gold price, the actual data to verify.3 the effectiveness of the proposed method is based on the real data) the price of gold, comparative analysis of the weighted moving average, exponential smoothing, grey model and ARIMA model on the prediction accuracy of the gold price trend and effectiveness, and summarized Summarizes the four models of the advantages and disadvantages and application scenarios. This paper mainly uses the grey model to predict the trend of gold price forecast based on ARIMA model of the gold price trend based on moving average model and exponential smoothing model, and predict the results of the model were analyzed. By comparing the effect of four kinds of model fitting analysis. Basically can describe the trend of gold prices, especially the ARIMA model fitting effect is the best. The actual data show that the model has a strong practical value in the short term forecast. Research and breakthroughs on the basis of this and other core technology, finally realizes the application of the whole system.

【学位授予单位】：国防科学技术大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F832.54;O211.61

【相似文献】