金融时间序列预测的数据挖掘方法比较以上证50指数为例
关键词 上证50指数 ARMA模型 支持向量机
毕业论文外文摘要
Title A comparative study of Financial Time Series Forecasting based on Data Mining: illustrated by the case of SSE 50 index
Abstract
As we all know, the stock market vagaries, is a complex system which is difficult to predict. The stock price is affected by many factors, such as political, economic, psychological and other investors, which makes the stock market price changes randomly varying characteristics. And because the stock market is an important part of a country's financial system, which can reflect the economic, political and social situation of a country to a certain extent, the stock price forecast has been an important issue. Securities prices based on time-varying characteristics, the researchers predict research on security prices largely established via statistical hypothesis traditionally autoregressive moving average model ARMA and auto-regressive integrated moving average model ARIMA model study. The last decade with the development of computer technology, combined with multi-disciplinary knowledge of data mining technology has gradually been used to predict the stock market in the analysis. So this paper will excavate the SSE 50 Index to predict technical support vector machine SVM ARMA model through traditional and popular data, and analysis of the traditional ARMA model and data mining technology advantages and disadvantages in terms of stock price prediction.
Keywords: ARMA SSE 50 Index support vector machine
目 次
1 引言 1
1.1研究背景1
1.2 证券预测方法1
1.3数据挖掘2
2 AMRA模型在证券价格预测的应用.5
2.1平稳时间序列5
2.2 ARMA 模型 7
2.3 ARMA 建模过程 13
2.4实证研究20
3. SVM支持向量机在证券价格预测的应用27
3.1 统计学习理论27
3.2支持向量机 32
3.3上证50指数的支持向量回归机预测 39
3.3 总结39
4.结论与总结结论 41
参考文献41
1 绪论
1.1研究背景
证券市场是一个国家或地区经济的重要组成部分，反映了一个国家或地区的经济运行状况。因此对于证券市场的价格预测受到了许多领域的研究者的关注，但是因为证券价格的变动受到了许多因素的影响，使得证券价格的精确预测十分困难，因此在早期证券分析人员往往只是对证券价格趋势进行简单的分析而不是直接预测具体的价格。随着20世纪20年代计量经济学的兴起，以ARMA模型为基础的时间序列预测方法得到了广泛的运用，对于证券价格的预测也有了较为精确的办法，但是因为ARMA模型是建立在平稳线性时间序列的基础上，对于非平稳非线性时间序列并不能得到很好的预测结果。而随着近几十年计算机技术的快速发展，基于数据挖掘技术的证券价格预测也越来越受到人们的注意。虽然已经有众多的研究者投身于这一领域的研究之中，并且出现了许多的数据挖掘技术的证券价格预测的方法，但是数据挖掘在证券价格预测方面是否比传统的时间序列模型分析方法更好仍然未得到解答。因此在本文中将对这两种方法进行实证比较研究。