多元统计分析

学分：3学分

先修课程：线性代数、统计学、R语言基础

课程简介：

本模块以多元统计方法为主，具体包括：多元回归、判别分析、聚类分析、主成分分析、因子分析、对应分析、典型相关分析和多维标度方法。在多元数据图形表示方面，将展示星相图、脸谱图、调和曲线图、矩阵散点图等，及均值向量、协方差矩阵等统计量。同时，矩阵运算、多元分布和数据变换内容作为基础简述。以上所有方法的实现均是基于R编程和统计软件包进行。

课程目标：

通过模块学习，学生应达到：运用多元图形和统计量对数据进行探索性分析，掌握多元统计模型方法，了解模型条件对应用的影响。针对同一类模型中具体算法差异进行比较，及多种模型之间的比较分析。根据模型编写R语言程序，并分析结果。使学生能够运用多元统计方法和软件对实际数据进行建模、编程与分析。

考核方式：

平时成绩（考勤、作业、分组讨论等）20%，论文报告30%，期末考试成绩50%

课程教材：

1.Brian Everitt, TorstenHothorn, An Introduction to Applied Multivariate Analysis with R. 2011, ISBN: 978-1441996497.

2.R 在网址http://www.r-project.org/下载

Multivariate Statistical Analysis

Credits: 3 Credits

Pre-requisite: Linear Algebra, Statistics, R Introduction

Module Description：

In this module, the cores are multivariate statistical methods. The main contents are: multiple regression, discriminant analysis, cluster analysis, principal component analysis, factor analysis, correspondence analysis, canonical correlation analysis and multidimensional scaling methods. Star plot, Chernoff-Flury Faces, Andrews’ Curves, scatterplot matrix will give representation of multivariate data, and statistics, such as mean vector, covariance matrix will be calculated. Meanwhile, introduce matrix algebra, multivariate distributions and data transformation for brief. Above methods in the module are based on the R programming and statistical package.

Learning Outcome and Objectives：

Through the module, students should be able to: give exploratory data analysis by multiple graphs and statistics, grasp method of multivariate statistical model, and understand the impact of model assumption on application. Compare algorithm differences between the same class model and various models. Write R programs for the model, and analyze the results. Then students can use multivariate statistical methods and software for real data modeling, programming and analysis.

Assessment：

Assignments and Class Performance  20%

Paper Report 30%

Final Exam 50%

Textbook and Reference：

1.Brian Everitt, TorstenHothorn, An Introduction to Applied Multivariate Analysis with R. 2011, ISBN: 978-1441996497.

2.R can be downloaded from http://www.r-project.org/