tp Analyse statistique avec R
vect <- seq(3, 12); sample(vect, 5)
length(vect)

sample(vect, 5, replace = TRUE)
sample(vect)
sample(10, 5)
sample(seq(1, 4), 3, prob = c(0.1, 0.2, 0.3, 0.4), replace = TRUE)
rnorm(100, 0, 1)
rnorm(100)
qnorm(0.95, 0, 1)
dnorm(0.5, 0, 1)
dnorm(2, 0, 1)

runif(n, min=0, max=1)
runif(100, 0, 5)   
runif(100, min= 0, max = 5)
a=rnorm(20,mean=55,sd=10)
mean(a)
sd(a)
max(a)
summary(a)
hist(a)
boxplot(a)
x1=rnorm(10,mean=100,sd=10)
x2=rnorm(10,mean=110,sd=10)
boxplot(x1,x2)
plot(x1,x2)
s1=rnorm(10,mean=2)
summary(s1)
s2=rnorm(100,mean=2)
summary(s2)
s3=rnorm(10000,mean=2)
summary(s3)
par(mfrow=c(3,3)) # organisation des graphiques selon une matrice 3 x 3
hist(s1) # histogrammes
hist(s2)
hist(s3)
X11()
plot(density(s1)) # fonction de densité
x=seq(-5,5,by=.01) # vecteur de coordonnées normales pour les abscisses
lines(x,dnorm(x,mean=2),col=2)
plot(density(s2))
lines(x,dnorm(x,mean=2),col=2)
plot(density(s3))
lines(x,dnorm(x,mean=2),col=2)
régression
x <- c(4, 6, 3, 5, 1, 9)
y <- 2 * x + 1 + rnorm(6, 0, 0.3)
lm(y ~x)
plot(y ~x); abline(lm(y ~x)).
lin$coefficients         
lin$fitted.values
lin$residuals  



summary(lin)             
predict(lin, newdata = data.frame(x = c(1, 2, 3, 4, 5)))
lin <- lm(z ~x + y)
lin <- lm(z ~x + y - 1)
predict(lin, data.frame(x = c(5, 7), y = c(3, 1)))
------------------------------------------------------------------------------
--------------------------------------------------------------------------------------------------