You are on page 1of 4

Note # 3

Objectives:

1. To introduce a number of probability distributions in R


2. To calculate probability in various probability distribution using R

Task # 1

In class we learnt about PDF and CDF

> help (Distributions) # list all available distributions in R

We will discuss: normal(norm), Student-t (t), Khi-Kuasa dua(chisq), F(f)

Four useful commands in R

Normal Distribution

> dnorm() # gives PDF

> pnorm() # gives CDF

> qnorm() # gives standard normal variate Z or ordinary normal X given a


probability

> rnorm() # general normally distributed random numbers

Example:

> dnorm(0)
[1] 0.3989423

The height of pdf at z=0 is 0.3989423.

> z=seq(-3,3,by=0.05) # gives z values from -3.0 to 3.0 with 0.05 increment

> y=dnorm(z) # gives heights of z values from -3.0 to 3.

> plot(z,y) # plotting z vs. y.

>pnorm(Z) # gives CDF

> args(pnorm) # gives you syntax on how to use this function


> args(pnorm)
function (q, mean = 0, sd = 1, lower.tail = TRUE, log.p = FALSE)

lower.tail = TRUE mean the probability is summed from left of Z whereas


lower.tail = FALSE means probability is summed from right of Z.

Try to get these values:

i) p(-∞ < z <=0)


ii) p(+∞ > z >=0)
iii) p(z < = 0.5)
iv) p( z > = 0.5)
v) p ( X > 3.5) if X ~ N(3, 1.5) (use two methods 1) directly using
pnorm() of non-standard normal and pnorm() of standard normal)

vi) p( >= 4.5 ) if a sample of size n=15 is taken from a population


of N(5, 1.5) (use two methods 1) directly using pnorm() of non-
standard normal and pnorm() of standard normal)

> qnorm(0.5) # gives Zo so that p(-∞ < z < Zo) =0.5. In this case Zo=0.
[1] 0

> qnorm(0.75) # 3 rd quantile

> qnorm(0.25) # 1st quantile

Try to get values of Zo or Xo from:

i) p( - ∞ < z < Zo)=0.7


ii) p( Zo < z < +∞ ) = 0.42
iii) p (- ∞ < X < Xo)=0.7 jika X ~ N(3, 1.5)

vii) p( >= Xo)= 0.75 if a sample of size n=15 is taken from N(5, 1.5)
(use two methods 1) directly using pnorm() of non-standard
normal and pnorm() of standard normal)

> zsample=rnorm(100) # generate random sample of saiz 100 from standard


normal
> hist(zsample) # plot the histogram

> xsample=rnorm(100, mean=5, sd=1.5) # generate random sample of size n=


100 from normal distribution of mean=5 and std=1.5)

> hist(xsample) # historgram a sample from normal distribution of mean=5 dan


sd=1.5
Student t Distribution

> dt(0,df=3) # PDF of t-dist with df=3

Compe t-distribution with different values of df e.g. df=10, 20, 50, 100 and
standard normal

> x=seq(-5,5,by=0.1) # generate values [-5:-1:5]


> y=dt(x,df=3) # PDF for t with df=3
> y1= dt(x,df=10)
> y2= dt(x,df=20)
> y3= dt(x,df=50)
> y4= dt(x,df=100)
> y5= dnorm(x)

> plot(x,y);lines(x,y) # plot x vs y and plot a line.

> lines(x,y1,col=’red’) # plot values of x vs y1 in same plot with red line


> lines(x,y2,col='green')
> lines(x,y3,col='blue')
> lines(x,y4,col='yellow')
> lines(x,y5,col='magenta')

Compares those different t-distribution and also with standard normal.

> pt(0,df=3) # CDF

> pt(0, df=3, lower.tail=TRUE)

> pt(0.5, df=3, lower.tail=TRUE)

Find these values:

i) p(-∞ < t <=0) for t(df=10)


ii) p(+∞ > t >=0) for t(df=10)
iii) p(t < = 0.5), for t(df=10)
iv) p( t > = 0.5), for t(df=10)
v) p ( X > 3.5) if X ~ N(3, σ 2), a sample of n= 20 taken with sample
variance of s2=1.2325

> qt(0.5, df=3) # What is the value of this?

> qt(0.25, df=3)

> qt (0.75, df=3)


> r=rt(100,df=20) # what is the value of this?

> hist(r)

Students can try chisq and f distribution.

You might also like