基于Matlab的数据多元回归分析的研究.docVIP

PAGE 36 - 摘 要 多元线性回归是利用MATLAB软件研究一个变量与多个变量的定量关系，MATLAB（矩阵实验室，是MATrix LABoratory的缩写）是一套高性能的数值运算和可视化软件，它集矩阵运算、数值分析、信号处理和图形显示于一体，构成了一个界面友好、使用方便的用户环境，是实现数据分析与处理的有效工具，其中MATLAB统计工具箱更为人们提供了一个强有力的数据统计分析工具。利用MATLAB统计工具箱来进行数据的多元回归分析使得分析的样本容量扩大，增加了统计推断的正确性，也促进了包含大量计算的多元统计分析的发展和运用。 本课题研究了在MATLAB软件平台上实现数据的多元统计分析，具体包括一元线性回归分析，非线性回归分析，多元线性回归分析，通过对基础数据分析函数polyfit（一元回归）；regress（多元回归）；及nlinfit（非线性回归）的学习。根据已得的实验结果以及以往的经验来建立统计模型，并研究变量之间的相关关系，建立起变量之间关系的近似表达式，并由此对相应的变量进行预测和控制。 根据所收集的数据，通过本文的研究方法进行一一分析，掌握它们的相关关系，可以找出数据中我们最需要的信息，从而进一步对总体的特性进行进一步的判断，把握规律，并将研究结果广泛运用于各种实际应用的预测和判断之中。 关键词：polyfit,regress,置信区间，最小二乘估计 Abstract Multiple linear regression is a variable and multiple variables are studied by using MATLAB software quantitative relationship of MATLAB (MATrix LABoratory, is the abbreviation of MATrix LABoratory) is a high-performance numerical calculation and visualization software, it combines MATrix operations, numerical analysis, signal processing and graphic display at an organic whole, constitute a friendly interface, easy to use user environment， including the MATLAB statistic tool box more people provided a powerful data statistical analysis tools. Using MATLAB toolbox for data statistics makes analysis of the sample size of expansion, multiple regression analysis to increase the validity of the statistical inference, also promoted the contains a large amount of computing multivariate statistical analysis of the development and use. This topic is studied on MATLAB software platform to realize data of multivariate statistical analysis, including a yuan linear regression analysis, nonlinear regression analysis, multivariate linear regression analysis, based on basic data analysis function polyfit (unary regression); Regress (multiple regression); Learning and nlinfit (nonlinear regression), According to the experimental results have been and past experience to build a statistical model, and studies the correlation between variables, establish the approximate expressions of the