R语言统计应用开发实战指南

下载需积分: 9 | PDF格式 | 10.62MB | 更新于2024-07-21 | 39 浏览量 | 举报

"R Statistical Application Development by Example Beginner's Guide" 是一本由Prabhanjan Narayanachar Tattar编写的书籍，旨在帮助初学者从零开始了解并掌握R语言在统计应用开发中的使用。本书结合了统计概念和R软件，以实例驱动的方式教授读者如何进行数据分析。本书的核心知识点包括： 1. **数据特性理解**：读者首先会学习到数据的本质，包括数据的类型、结构和分布，这对于后续的数据分析至关重要。 2. **描述性统计与探索性分析**：书中强调了探索性数据分析（EDA）的重要性，它超越了简单的数据汇总，通过可视化工具和统计测试帮助读者发现数据中的模式和趋势。 3. **统计推断**：讲解了基于样本数据对总体参数进行估计和假设检验的基本原理，利用R语言的计算能力进行模拟，使读者能理解和应用统计推断方法。 4. **回归建模**：详细介绍了线性回归模型，包括模型的建立、参数估计、误差分析以及模型诊断。此外，还涵盖了逻辑回归（用于处理分类问题）和决策树算法（CART），这些是解决实际复杂问题的重要工具。 5. **高级统计模型**：书中的内容还包括了更现代和先进的统计模型，如CART决策树，以及空间统计，这些都是处理具有空间相关性的数据的关键方法。 6. **R语言实践**：每一步教学都与实际的数据和R代码相结合，让读者能够通过实践来深化理解，提高应用能力。 7. **学习方法**：作者采用清晰和教学导向的方式来编写，确保读者能够在实例中逐步掌握R语言在统计应用开发中的技巧。通过本书的学习，读者不仅能够建立起坚实的统计基础，还能熟练运用R语言处理各种实际问题，从数据预处理到模型构建，再到结果解释，都能得心应手。无论你是统计学的新手还是寻求提升R编程技能的实践者，这本书都将提供一条有效且有趣的路径。

Table of Contents

[ v ]

Time for acon – residual plots for the logisc regression model 214

Inuence and leverage for the GLM 216

Time for acon – diagnoscs for the logisc regression 216

Receiving operator curves 220

Time for acon – ROC construcon 221

Logisc regression for the German credit screening dataset 223

Time for acon – logisc regression for the German credit dataset 225

Summary 227

Chapter 8: Regression Models with Regularizaon 229

The overng problem 230

Time for acon – understanding overng 231

Regression spline 234

Basis funcons 234

Piecewise linear regression model 235

Time for acon – ng piecewise linear regression models 235

Natural cubic splines and the general B-splines 238

Time for acon – ng the spline regression models 239

Ridge regression for linear models 243

Time for acon – ridge regression for the linear regression model 244

Ridge regression for logisc regression models 248

Time for acon – ridge regression for the logisc regression model 248

Another look at model assessment 250

Time for acon – selecng lambda iteravely and other topics 250

Summary 256

Chapter 9: Classicaon and Regression Trees 257

Recursive parons 258

Time for acon – paroning the display plot 259

Spling the data 260

The rst tree 260

Time for acon – building our rst tree 261

The construcon of a regression tree 265

Time for acon – the construcon of a regression tree 266

The construcon of a classicaon tree 276

Time for acon – the construcon of a classicaon tree 279

Classicaon tree for the German credit data 284

Time for acon – the construcon of a classicaon tree 284

Pruning and other ner aspects of a tree 288

Time for acon – pruning a classicaon tree 289

Summary 291

Preface

The open source soware R is fast becoming one of the preferred companions of Stascs

even as the subject connues to add many friends in Machine Learning, Data Mining, and so

on among its already rich scienc network. The era of mathemacal theory and stascal

applicaons embeddedness is truly a remarkable one for the society and the soware has

played a very pivotal role in it. This book is a humble aempt at presenng Stascal Models

through R for any reader who has a bit of familiarity with the subject. In my experience of

praccing the subject with colleagues and friends from dierent backgrounds, I realized that

many are interested in learning the subject and applying it in their domain which enables

them to take appropriate decisions in analyses, which involves uncertainty. A decade earlier

my friends would be content with being pointed to a useful reference book. Not so anymore!

The work in almost every domain is done through computers and naturally they do have

their data available in spreadsheets, databases, and somemes in plain text format. The

request for an appropriate stascal model is invariantly followed by a one word queson

"Soware?" My answer to them has always been a single leer reply "R!" Why? It is really a

very simple decision and it has been my companion over the last seven years. In this book,

this experience has been converted into detailed chapters and a cleaner breakup of model

building in R.

A by-product of the interacon with colleagues and friends who are all aspiring stascal

model builders has been that I have been able to pick up the trough of their learning

curve of the subject. The rst aempt towards xing the hurdle has been to introduce

the fundamental concepts that the beginners are most familiar with, which is data.

The dierence is simply in the subtlees and as such I rmly believe that introducing the

subject on their turf movates the reader for a long way in their journey. As with most

stascal soware, R provides modules and packages which mostly cover many of the

recently invented stascal methodology. The rst ve chapters of the book focus on the

fundamental aspects of the subject and the R soware and hence cover R basics, data

visualizaon, exploratory data analysis, and stascal inference.

Preface

[ 2 ]

The foundaonal aspects are illustrated using interesng examples and sets up the

framework for the later ve chapters. Regression models, linear and logisc regression

models being at the forefront, are of paramount interest in applicaons. The discussion is

more generic in nature and the techniques can be easily adapted across dierent domains.

The last two chapters have been inspired by the Breiman school and hence the modern

method of Classicaon and Regression Trees has been developed in detail and illustrated

through a praccal dataset.

What this book covers

Chapter 1, Data Characteriscs, introduces the dierent types of data through a

quesonnaire and dataset. The need of stascal models is elaborated in some interesng

contexts. This is followed by a brief explanaon of R installaon and the related packages.

Discrete and connuous random variables are discussed through introductory R programs.

Chapter 2, Import/Export Data, begins with a concise development of R basics. Data frames,

vectors, matrices, and lists are discussed with clear and simpler examples. Imporng of data

from external les in csv, xls, and other formats is elaborated next. Wring data/objects from

R for other soware is considered and the chapter concludes with a dialogue on R session

management.

Chapter 3, Data Visualizaon, discusses ecient graphics separately for categorical and

numeric datasets. This translates into techniques of bar chart, dot chart, spine and mosaic

plot, and four fold plot for categorical data while histogram, box plot, and scaer plot for

connuous/numeric data. A very brief introducon to ggplot2 is also provided here.

Chapter 4, Exploratory Analysis, encompasses highly intuive techniques for preliminary

analysis of data. The visualizing techniques of EDA such as stem-and-leaf, leer values, and

modeling techniques of resistant line, smoothing data, and median polish give a rich insight

as a preliminary analysis step.

Chapter 5, Stascal Inference, begins with the emphasis of likelihood funcon and

compung the maximum likelihood esmate. Condence intervals for the parameters

of interest is developed using funcons dened for specic problems. The chapter also

considers important stascal tests of Z-test and t-test for comparison of means and

chi-square tests and F-test for comparison of variances.

Chapter 6, Linear Regression Analysis, builds a linear relaonship between an output and a

set of explanatory variables. The linear regression model has many underlying assumpons

and such details are veried using validaon techniques. A model may be aected by a

single observaon, or a single output value, or an explanatory variable. Stascal metrics

are discussed in depth which helps remove one or more kinds of anomalies. Given a large

number of covariates, the ecient model is developed using model selecon techniques.

Preface

[ 3 ]

Chapter 7, The Logisc Regression Model, is useful as a classicaon model when the output

is a binary variable. Diagnosc and model validaon through residuals are used which lead

to an improved model. ROC curves are next discussed which helps in idenfying of a beer

classicaon model.

Chapter 8, Regression Models with Regularizaon, discusses the problem of over ng

arising from the use of models developed in the previous two chapters. Ridge regression

signicantly reduces the possibility of an over t model and the development of natural

spine models also lays the basis for the models considered in the next chapter.

Chapter 9, Classicaon and Regression Trees, provides a tree-based regression model.

The trees are inially built using R funcons and the nal trees are also reproduced using

rudimentary codes leading to a clear understanding of the CART mechanism.

Chapter 10, CART and Beyond, considers two enhancements of CART using bagging and

random forests. A consolidaon of all the models from Chapter 6 to Chapter 10 is also given

through a dataset.

Chapter 1 to Chapter 5 form the basics of R soware and the Stascs subject. Praccal

and modern regression models are discussed in depth from Chapter 6 to Chapter 10.

Appendix, References, lists names of the books that have been referred to in this book.

What you need for this book

R is the only required soware for this book and you can download it from http://www.

cran.r-project.org/

. R packages will be required too though this task is done within a

working R session. The datasets used in the book is available in the R package RSADBE, which

is an abbreviaon of the book's tle, at http://cran.r-project.org/web/packages/

RSADBE/index.html

Who this book is for

This book will be useful for readers who have air and a need for stascal applicaons

in their own domains. The rst seven chapters are also useful for any masters students in

Stascs and the movated student can easily complete the rest of the book and obtain

a working knowledge of CART.

Conventions

In this book, you will nd several headings appearing frequently.

To give clear instrucons of how to complete a procedure or task, we use:

www.allitebooks.com

剩余344页未读，继续阅读

vanridin

粉丝: 108

R语言统计应用开发实战指南

R Statistical Application Development by Example Beginner's Guide

Statistical Application Development with R and Python, 2nd-Packt (2017).pdf

Statistical Application Development R Python

Statistical Application Development with R and Python(2nd) mobi

Statistical Application Development with R and Python(2nd) epub

Statistical Application Development with R and Python(2nd) azw3

Statistical Application Development with R and Python(2nd) 无水印原版pdf

Statistical Analysis with R -- Beginner's Guide.pdf

A Beginner's Guide to R

Numpy Beginner's Guide - Third Edition

最新资源