Sunteți pe pagina 1din 38

9 October 2019

Topic 7:
Distribution of Sampling Statistics

FASILKOM, Universitas Indonesia


CSF2600102 – Statistics and Probability

Fakultas Ilmu Komputer


Universitas Indonesia
Ganjil 2015/2016
1
Outline
• Distributions Arising from the Normal
• The Chi-Square Distribution

9 October 2019
• The t-Distribution
• The F-Distribution
• Introduction to Distribution of Sampling Statistics
• Sample Mean

FASILKOM, Universitas Indonesia


• Central Limit Theorem
• Approximate Distribution of Sample Mean
• How Large a Sample is Needed
• Sample Variance
• Sampling Distribution from Normal Distribution
• Distribution of Sample Mean
• Joint Distribution of Sample Mean and Variance 2
• Sampling from a Finite Population
the Normal
Distributions Arising from

FASILKOM, Universitas Indonesia 9 October 2019


3
Various Distributions Arising from
The Normal

9 October 2019
1. The Chi-Square Distribution
2. The t-Distribution

FASILKOM, Universitas Indonesia


3. The F-Distribution

4
1. Chi-Square Distribution

FASILKOM, Universitas Indonesia 9 October 2019


5
1. Chi-Square Distribution
• If Z1, Z2 , Z3, …, Zn are independent standard

9 October 2019
normal distribution variables, then X, defined
by:
X= Z12 + Z22 + Z32 + …+ Zn2

FASILKOM, Universitas Indonesia


is said to have a Chi-Square distribution with
n degree of freedom.
• Notation:

6
Chi-Square Density Function
• If X is a chi-square random variable with

9 October 2019
n degrees of freedom, then for any α ϵ
(0,1): the quantity χ2α,n is defined to be
such that:

FASILKOM, Universitas Indonesia


P( X   ,n )  
2

7
FASILKOM, Universitas Indonesia 9 October 2019
8
Degree of Freedom
• In statistics, the number of degrees of freedom is the
number of values in the final calculation of a statistic

9 October 2019
that are free to vary
• For example, if we have two observations:
• when calculating the mean we have two

FASILKOM, Universitas Indonesia


independent observations;
• however, when calculating the variance, we have
only one independent observation, since the two
observations are equally distant from the mean.

9
Example
• Suppose that we are attempting to

9 October 2019
locate a target in three-dimensional
space, and that the three coordinate
errors (in meters) of the point chosen

FASILKOM, Universitas Indonesia


are independent normal random
variables with mean 0 and standard
deviation 2. Find the probability that
the distance between the point chosen
and the target exceeds 3 meters. 10
FASILKOM, Universitas Indonesia 9 October 2019
11
2. The t-Distribution

FASILKOM, Universitas Indonesia 9 October 2019


12
2. The t-Distribution
 2
• If Z and n are independent random variables, with Z
having a standard normal distribution and  n having
2

9 October 2019
a chi-square distribution with n degrees of freedom,
then the random variable Tn defined by

FASILKOM, Universitas Indonesia


is said to have a t-distribution with n degrees of
freedom.

13
9 October 2019
FASILKOM, Universitas Indonesia
Like the standard normal density, the t-density
is symmetric about zero. In addition, as n
becomes larger, it becomes more and more like
a standard normal density. 14
9 October 2019
FASILKOM, Universitas Indonesia
Notice that the t -density has thicker “tails,”
indicating greater variability, than does the
normal density. 15
3. The F-Distribution

FASILKOM, Universitas Indonesia 9 October 2019


16
3. The F-Distribution
 
• If n and m are independent chi-square
2 2

9 October 2019
random variables with n and m degrees of
freedom, respectively, then the random
variable Fn,m defined by

FASILKOM, Universitas Indonesia


is said to have an F-distribution with n and
m degrees of freedom.
17
9 October 2019
FASILKOM, Universitas Indonesia
The quantities Fα,n,m are tabulated in Table A4 of
the Appendix for different values of n,m, and α ≤ 18
1/2
Introduction to Distribution of Sampling Statistics

Our objective in this chapter: to make inferences

9 October 2019
about a distribution F using the samples taken from F.

• Parametric inference problem: The form of F is

FASILKOM, Universitas Indonesia


specified up to a set of unknown parameters, e.g.:
• F is assumed as a normal distribution function having
an unknown mean and variance.

• Nonparametric inference problem: nothing is


assumed about the form of F. 19
Sample Mean
• Suppose X1,X2,...,Xn are the samples taken from any

9 October 2019
distribution whose population mean is µ and population
variance is 2.

FASILKOM, Universitas Indonesia


• The sample mean is

20
Sample Mean
X is also a random variable whose mean is given

9 October 2019
by:

FASILKOM, Universitas Indonesia


21
Sample Mean

The variance of X is given by:

FASILKOM, Universitas Indonesia 9 October 2019


22
Sample Mean
• X is also centered about the population mean μ, but its

9 October 2019
spread becomes more and more reduced as the sample
size increases.

FASILKOM, Universitas Indonesia


23
Sample Variance

9 October 2019
FASILKOM, Universitas Indonesia
 The expected value of sample variance is the population
variance 2

24
The Central Limit Theorem

Let X1,X2,...,Xn be a sequence of independent and

9 October 2019
identically distributed (i.i.d) random variables each
having mean μ and variance σ2. Then for n large,

FASILKOM, Universitas Indonesia


i.e., the distribution of X1+X2+...+Xn is approximately
normal with mean nμ and variance nσ2.
25
The Central Limit Theorem

FASILKOM, Universitas Indonesia 9 October 2019


26
Example
An insurance company has 25,000 automobile

9 October 2019
policy holders. If the yearly claim of a policy holder
is a random variable with mean 320 and standard

FASILKOM, Universitas Indonesia


deviation 540, approximate the probability that
the total yearly claim exceeds 8.3 million.

27
FASILKOM, Universitas Indonesia 9 October 2019
28
Approximate Distribution of the
Sample Mean

9 October 2019
Since the sample mean has expected value μ and
standard deviation σ/√n, it then follows that for n
large,

FASILKOM, Universitas Indonesia


has approximately a standard normal distribution.

29
Example
The weights of a population of workers have

9 October 2019
mean of 167 and standard deviation of 27.
(a) If a sample set of 36 workers is chosen,

FASILKOM, Universitas Indonesia


approximate the probability that the sample
mean of their weights lies between 163 and
171.
(b) Repeat part (a) when the sample is of size
144.
30
FASILKOM, Universitas Indonesia 9 October 2019
31
FASILKOM, Universitas Indonesia 9 October 2019
32
How Large a Sample Is Needed
• A general rule of thumb is that one can be

9 October 2019
confident of the normal approximation
whenever the sample size n is at least 30.

FASILKOM, Universitas Indonesia


• That is, practically speaking, no matter how non
normal the underlying population distribution
is, the sample mean of a sample of size at least
30 will be approximately normal.
• In most cases, the normal approximation is valid
33
for much smaller sample sizes.
FASILKOM, Universitas Indonesia 9 October 2019
34
Sampling from a Finite Population

• Consider a population of N elements, and suppose that:

9 October 2019
• p is the proportion of the population that has a certain
characteristic of interest;
• Np elements have this characteristic, and N(1− p) do not.

FASILKOM, Universitas Indonesia


• A sample of size n from this population is said to be a
random sample if it is chosen in such a manner that each
of the C(N,n) population subsets of size n is equally likely
to be the sample.

38
Sampling from A Finite Population
• Suppose n samples are chosen, i=1,...,n, let

9 October 2019
FASILKOM, Universitas Indonesia
• If N >> n, then X1,X2,... ,Xn are approximately
n
independent. Let X   X , then X is a
i 1
i

binomial R.V. with parameters n and p, and


E  X   np
39
Var  X   np 1- p  .
Sampling from a Finite Population

9 October 2019
• Let X  proportion of the sample that has the
n
characteristic, then

FASILKOM, Universitas Indonesia


E[ X ]
E[ X ]  p
n
1 p(1  p)
Var ( X )  2 [np(1  p)]  .
n n
40
The End of Topic 7

FASILKOM, Universitas Indonesia 9 October 2019


41

S-ar putea să vă placă și