摘 要
回归分析预测被广泛应用于经济学、社会科学、工程技术和生物学等许多领域当中，进行各专业指标变量的统计分析和预测控制。本文从回归分析预测与其他预测方法的简单算法对比出发，系统的讨论了线性回归分析和非线性回归分析的基本算法，再以八一钢铁股票的历史价格为例，对比多元线性回归和非线性回归分析预测，得出非线性回归分析拟合能力更强、拟合优度更高的结论。
关键字：回归分析预测；非线性回归；线性回归；拟合度
I
Regression analysis in the stock of the use of
forecast
Abstract: The regression analysis forecast is widely applied in the economic, the social sciences,
the engineering technology and the biology and so on the middle of many domains, carries on each specialized target variable the statistical analysis and the predictive control. This article embarks from the regression analysis forecast with other forecast technique's simple algorithm contrast, system's discussion linear regression analysis and non-linear regression analysis's primary algorithm, again take 81 steel and iron stock's historical price as the example, the contrast multi-dimensional linear regression and the non-linear regression analysis predict that obtained the non-linear regression analysis fitting ability to be stronger, a goodness of fit higher conclusion.
Keywords: Forecast regression analysis，Non-linear regression，Linear regression，Fit
II
目 录
第1章 前 言 ............................................................................................................ 1 1.1 选题背景和意义 ................................................................................................ 1 1.2 股票的可预测性 ................................................................................................ 2 1.3 回归的发展概况 ................................................................................................ 3 1.4 文章结构 ............................................................................................................ 4 第2章 预测方法概述 ................................................................................................ 6 2.1 趋势分析法 ........................................................................................................ 6 2.2 时间序列法 ........................................................................................................ 7 2.3 灰色预测法 ........................................................................................................ 7 2.4 模糊数学法 ........................................................................................................ 8 2.5 回归分析法 ........................................................................................................ 9 第3章 回归分析 ...................................................................................................... 10 3.1 线性回归分析 .................................................................................................. 11 3.1.1 一元回归模型 ........................................................................................... 11 3.1.2 多元回归线性分析模型 ........................................................................... 12 3.1.3 线性相关程度测定及相关性检验预测 ................................................... 16 3.2 非线性回归分析 .............................................................................................. 17 3.2.1 非线性回归分析 ....................................................................................... 17 3.2.2 参数估计和模型检验 ............................................................................... 19 3.2.3 非线性回归分析存在的问题 ................................................................... 21 第4章 实例分析 ...................................................................................................... 23 4.1 用回归分析进行预测的步骤 .......................................................................... 23 4.2 数据的选取 ...................................................................................................... 23 4.3 线性回归分析的程序实现 .............................................................................. 24 4.3.1 回归方程求解 ........................................................................................... 24 4.3.2 运行结果 ................................................................................................... 26
III
回归分析在股票预测中的应用
