cyrilobolonsky

#non-parametric tests in R
dengue.fewer <-c(3000,3200,3500,5068,5679,6200,6300,7020)
scrub.typhus<-c(4400,4500,5900,6839,7561,9047,12300,14000)

wilcox.test(dengue.fewer,scrub.typhus)
t.test(dengue.fewer,scrub.typhus,var.equal=T)


ttest.wilcox.examples <- function(x,y,z,k)
{

cyrilobolonsky

#Conducting t-tests in R.
t.test(death_rt)
t.test(lifeexpf, lifeexpm)
t.test(lifeexpf, lifeexpm, var.equal=T)
t.test(lifeexpf, lifeexpm, paired=T)

#A function for creating random examples of t-tests
ttest.for.examination <- function(x,y,z,k)
{
  subjects <- x

cyrilobolonsky

#Visualising Hypothesis Tests in R
#The red distribution is what you can expect to see if you plot repeated samples when the null hypothesis is true.
#You can recognize the Ho because it sounds like: "there was no difference", for instance: "The intervention did not affect the tumor marker."

#one tailed test   
#rare in health sciences, more common in industrial process control
x=seq(50,140,length=200)
y1=dnorm(x,80, 10)
plot(x,y1,type='l',lwd=2,col='red')
y2=dnorm(x,110, 10)

cyrilobolonsky

#Shapiro test shows whether the variables have a normal distribution
shapiro.test(lifeexpf)
shapiro.test(lifeexpm)
shapiro.test(birth_rt)
shapiro.test(death_rt)
shapiro.test(world[,2])
shapiro.test(world[,3])
shapiro.test(world[,5])
#We are alomst 100% sure that NONE of the variables follows a normal distribution.

cyrilobolonsky

#Probability Distributions in R


#BINOMINAL DISTRIBUTION
?dbinom
x <- 0:30  
#binominal - density
plot(x, dbinom(x, 30, 0.5), type = "h")
#binominal - cumulative distribution function
plot(x, pbinom(x, 30, 0.5), type = "h")

cyrilobolonsky

#Principal Component Analysis with R using the library FacoMineR
#1)preliminaries
#adding the file to R
auto<-read.table("auto2004.csv", header=TRUE, sep=",")
#viewing the structure of the file
str(auto)
#attaching the file to the R memory
attach(auto)
#installing the library FactoMineR
library(FactoMineR)