Book Image

R Statistical Application Development by Example Beginner's Guide

By : Prabhanjan Narayanachar Tattar
Book Image

R Statistical Application Development by Example Beginner's Guide

By: Prabhanjan Narayanachar Tattar

Overview of this book

<p>"R Statistical Application Development by Example Beginner’s Guide" explores statistical concepts and the R software, which are well integrated from the word go. This demarcates the separate learning of theory and applications and hence the title begins with “R Statistical …”. Almost every concept has an R code going with it which exemplifies the strength of R and applications. Thus, the reader first understands the data characteristics, descriptive statistics, and the exploratory attitude which gives the first firm footing of data analysis. Statistical inference and the use of simulation which makes use of the computational power complete the technical footing of statistical methods. Regression modeling, linear, logistic, and CART, builds the essential toolkit which helps the reader complete complex problems in the real world.<br /><br />The reader will begin with a brief understanding of the nature of data and end with modern and advanced statistical models like CART. Every step is taken with DATA and R code.<br /><br />The data analysis journey begins with exploratory analysis, which is more than simple descriptive data summaries, and then takes the traditional path up to linear regression modeling, and ends with logistic regression, CART, and spatial statistics.<br /><br />True to the title R Statistical Application Development by Example Beginner’s Guide, the reader will enjoy the examples and R software.</p>

Related Content you might be interested in

Current Title:

Small book image
R Statistical Application Development by Example Beginner's Guide
Prabhanjan Narayanachar Tattar
Jul, 2013 | 344 pages
No titles found
Table of Contents (18 chapters)
R Statistical Application Development by Example Beginner's Guide
R Statistical Application Development by Example Beginner's Guide
Credits
Credits
About the Author
About the Author
About the Reviewers
About the Reviewers
www.PacktPub.com
www.PacktPub.com
Preface
Preface
Free Chapter
1
Data Characteristics
Data Characteristics
Questionnaire and its components
Experiments with uncertainty in computer science
R installation
Continuous distribution
Summary
2
Import/Export Data
Import/Export Data
data.frame and other formats
Time for action – understanding constants, vectors, and basic arithmetic
Time for action – matrix computations
Time for action – creating a list object
Time for action – creating a data.frame object
Summary
3
Data Visualization
Data Visualization
Visualization techniques for categorical data
Time for action – bar charts in R
Time for action – dot charts in R
Time for action – the spine plot for the shift and operator data
Time for action – the mosaic plot for the Titanic dataset
Visualization techniques for continuous variable data
Time for action – using the boxplot
Time for action – understanding the effectiveness of histograms
Time for action – plot and pairs R functions
A brief peek at ggplot2
Time for action – qplot
Time for action – ggplot
Summary
4
Exploratory Analysis
Exploratory Analysis
Essential summary statistics
Time for action – the essential summary statistics for "The Wall" dataset
The stem-and-leaf plot
Time for action – the stem function in play
Letter values
Data re-expression
Bagplot – a bivariate boxplot
Time for action – the bagplot display for a multivariate dataset
The resistant line
Time for action – the resistant line as a first regression model
Smoothing data
Time for action – smoothening the cow temperature data
Median polish
Time for action – the median polish algorithm
Summary
5
Statistical Inference
Statistical Inference
Maximum likelihood estimator
Time for action – visualizing the likelihood function
Time for action – finding the MLE using mle and fitdistr functions
Confidence intervals
Time for action – confidence intervals
Hypotheses testing
Time for action – testing the probability of success
Time for action – testing proportions
Time for action – testing one-sample hypotheses
Time for action – testing two-sample hypotheses
Summary
6
Linear Regression Analysis
Linear Regression Analysis
The simple linear regression model
Time for action – the arbitrary choice of parameters
Time for action – building a simple linear regression model
Time for action – ANOVA and the confidence intervals
Time for action – residual plots for model validation
Multiple linear regression model
Time for action – averaging k simple linear regression models
Time for action – building a multiple linear regression model
Time for action – the ANOVA and confidence intervals for the multiple linear regression model
Time for action – residual plots for the multiple linear regression model
Regression diagnostics
The multicollinearity problem
Time for action – addressing the multicollinearity problem for the Gasoline data
Model selection
Time for action – model selection using the backward, forward, and AIC criteria
Summary
7
The Logistic Regression Model
The Logistic Regression Model
The binary regression problem
Time for action – limitations of linear regression models
Probit regression model
Time for action – understanding the constants
Logistic regression model
Time for action – fitting the logistic regression model
Time for action – The Hosmer-Lemeshow goodness-of-fit statistic
Model validation and diagnostics
Time for action – residual plots for the logistic regression model
Time for action – diagnostics for the logistic regression
Receiving operator curves
Time for action – ROC construction
Logistic regression for the German credit screening dataset
Time for action – logistic regression for the German credit dataset
Summary
8
Regression Models with Regularization
Regression Models with Regularization
The overfitting problem
Time for action – understanding overfitting
Regression spline
Time for action – fitting piecewise linear regression models
Time for action – fitting the spline regression models
Ridge regression for linear models
Time for action – ridge regression for the linear regression model
Ridge regression for logistic regression models
Time for action – ridge regression for the logistic regression model
Another look at model assessment
Time for action – selecting lambda iteratively and other topics
Summary
9
Classification and Regression Trees
Classification and Regression Trees
Recursive partitions
Time for action – partitioning the display plot
Time for action – building our first tree
The construction of a regression tree
Time for action – the construction of a regression tree
The construction of a classification tree
Time for action – the construction of a classification tree
Classification tree for the German credit data
Time for action – the construction of a classification tree
Pruning and other finer aspects of a tree
Time for action – pruning a classification tree
Summary
10
CART and Beyond
CART and Beyond
Improving CART
Time for action – cross-validation predictions
Bagging
Time for action – understanding the bootstrap technique
Time for action – the bagging algorithm
Random forests
Time for action – random forests for the German credit data
The consolidation
Time for action – random forests for the low birth weight data
Summary
References
References
Index
Index

Time for action – qplot

Here, we will first use the qplot function for obtaining various kinds of plots. To keep the story short, we are using the earlier datasets only and hence, a reproduction of the similar plots using qplot won't be displayed. The reader is encouraged to check ?qplot and its examples.

  1. Load the library with library(ggplot2).

  2. Rearrange the resistivity dataset in a different format and obtain the boxplots:

    test <- data.frame(rep(c("R1","R2"),each=8),c(resistivity[,1], resistivity[,2]))
names(test) <- c("RES","VALUE")
qplot(factor(RES),VALUE,data=test,geom="boxplot")

    The gplot function needs to be explicitly specified that RES is a factor variable and according to its levels, we need to obtain the boxplot for the resistivity values.

  3. For the Gasoline dataset, we would like to obtain a boxplot of the mileage accordingly as the gear system could be manual or automatic. Thus, the qplot can be put to action with qplot(factor(x11),y,data=Gasoline,geom= "boxplot").

  4. A histogram...

Previous Section
End of Section 12
Next Section