Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

MULTINOMIAL DISTRIBUTION

Definition:
The multinomial distribution is the multivariate analogue of the Binomial
distribution. It is one of the simplest but most important discrete multivariate
distributions.
The binomial arises from a random experiment in which a finite sequence of n
repeated and independent Bernoulli trials are conducted, each trial resulting
in only one or two possible outcomes, success or failure. The multinomial
distribution arises when the number of possible outcomes of a single trial is
generated from 2 to k.
Let 𝐸1 , 𝐸2 , … , 𝐸𝑘 denote the k possible outcomes in a single trial of the
experiment. Let n repeated trials be conducted and the various outcomes noted.
Let 𝑋𝑖 denote the number of times the outcomes 𝐸𝑖 is observed with
corresponding probabilities 𝑝𝑖 (𝑖 = 1,2, … , 𝑘). As the outcomes are mutually
exclusive.
𝑋1 + 𝑋2 + ⋯ + 𝑋𝑘 = 𝑛
𝑎𝑛𝑑, 𝑝1 + 𝑝2 + ⋯ + 𝑝𝑘 = 1

pmf;
Example:
Consider the production of ball bearings of a certain type, the diameter of which
should be 0.25 inch. Because of the inherent variability in the manufacturing
process, the bearings are classified as ‘undersize’, ’oversize’ and ‘acceptable’
if they measure less than 0.2495 inch, more than 0.2505 inch and between
0.2495 and 0.2505 inch, respectively. Suppose the production process is such
that 4% of the bearings are undersize, 6% are oversize and 90% are acceptable.
If 100 of these are picked at random, what is the probability of getting 𝑥1
undersize, 𝑥2 oversize, and 𝑥3 acceptable bearings?
Solution: Let, 𝑋1 represent undersize bolt, 𝑋2 =Oversize and
𝑋3 =Acceptable bolts.
Now, n=100, 𝑝1 =0.04, 𝑝2 =0.06, 𝑝3 =0.90 The required probability is
given by-
100!
𝑃(𝑥1 , 𝑥2 , 𝑥3 ) =
(𝑥1 ! 𝑥2 ! 𝑥3 ! (0.04) 𝑥1 (0.06) 𝑥2 (0.09)𝑥3
𝑤ℎ𝑒𝑟𝑒, 0 ≤ 𝑥𝑖 ≤ 100; 𝑖 = 1,2,3 𝑎𝑛𝑑 ∑ 𝑥𝑖 = 100.

Example:
Suppose 6 bearings are sampled from the process.
1. What is the prob that there will be two bearings of each category?
2. What is the prob that all of them will be accepted?
Solution:
1. 𝐺𝑖𝑣𝑒𝑛, 𝑛 = 6, 𝑥1 = 2, 𝑥2 = 2, 𝑥3 = 2. The required probability is given
6!
by = ( )( )2 ( )2 ( )2
=0.00041999.
2!2!2! 0.04 0.06 0.09

2. 𝐺𝑖𝑣𝑒𝑛, 𝑛 = 6, 𝑥1 = 0, 𝑥2 = 0, 𝑥3 = 6. The required probability is given


6!
by = ( )( )0 ( )0 ( )6
= 0.5314.
0!0!6! 0.04 0.06 0.09
Home Work:
1. Write down the properties of binomial distribution?
2. What are the properties of multinomial distribution?
Variance:

Home Work:
3. Find third and fourth central moment of multinomial distribution.
MLE of Multinomial Distribution:

Frequency Distribution of Multinomial Distribution


Lagrange’s Multiplier:
Multinomial Calculator:
http://stattrek.com/online-calculator/multinomial.as

Binomial Theorem:

You might also like