Discrete Random Variable with Probability Bayess Theorem
Mass Function P ( A|B ) P ( B )
P ( B| A )=
w 0 1 2 3 4 5 P ( B ) P ( A|B ) + P ( B' ) P( AB' )
0.4 a b 0.10 0.05 0.05 Complement Rule for Conditional Probabilities
f W ( w )=P(W =w)
P ( A '|B ) =1P( AB)
E ( W )=00.4 +1a+2b +30.10+40.05+ 50.0 5=1.3
65% of all women who submit to pregnancy test are
p .m . f =0.4+a+ b+0.1+0.05+ 0.05=1 actually pregnant. A certain test gives false positive
results with probability 0.02 and valid positive result
Solve two equationsunknowns find a=0.25b=0.15with probability 0.99
Among women who submit to a pregnancy
Moment Generating Function for W test, what fraction of the tests is positive?
tX t 2t 3t 4t Let 5T=test positive and A=actually pregnant,
M ( t )=E e =0.4+ 0.25 e + 0.15 e +0.1 e +0.05 e + 0.05 et
Then
P ( A )=0.65, P ( T A )=0.99, P ( T | A' )=0.0 2
Stan dard Deviation of W ( W )
2 2 P (T )=P ( T A )+ P ( T A ' )
Var ( W )=E ( W ) [ E ( W ) ] =
2 2 2 2 2P (T 2)=P ( A )P ( T | A ) + P ( A' )P (T A ' )
2
0 0.4+1 0.25+2 0.15+3 0.10+ 4 0.05+5 0.051.3
P (T )=0.650.99+ 0.350.02=0.6506=65.05
W = Var (W )= 2.11=1.453
If a womans test is positive, what is the
probability that she is actually pregnant?
Binomial Distribution P (T A ) P ( T| A ) P ( A ) 0.650.99
The probability that a circuit board coming off an P ( A|T ) = = = =0.9892
assembly line needs rework is 0.15 Suppose 12 boards are P(T ) P (T ) 0.6505
tested and all boards are independent of each other
Probability that EXACTLY four will need rework If a particular womans test is negative, what
X b(12, 0.15) is the probability that she is actually
pregnant?
P ( A T ' ) P ( T '| A ) P ( A ) ( 1P ( T| A ) ) P ( A )
P ( X=4 ) = 12 0.1540.858=0.06828
4 ( ) P ( A|T ' ) =
P( T' )
=
P(T )
'
=
1P ( T )
Probability that AT LEAST one needs rework INDEPENDENCE
12
P ( X 1 )=1P ( X=0 )=10.85 =0.8578 A and B are independent if they satisfy the
product formula
Probability that AT MOST two needs rework
P ( A B ) =P ( A ) P ( B )P ( B A ) =P ( B ) P( A )
P ( X 2 )=P ( X =0 ) + P ( X =1 )+ P ( X =2 )=
Independence of Compliments:
12 0.1500.8512 + 12 0.1510.8511 + 12 0.15 20.8510 If
( )
5 1 ( ) 2 ( ) =A and B are independent, so are A B, A B
and A B
Connection between independence and
0.1422+0.3012+0.2924=0.7358 conditional prob
Conditional Probability & Bayes Theorem P ( A|B )=P ( A ) iff AB are independent
The Conditional Probability of event A given
event B is: Pairwise Independent events A, B, and C
P ( A B) P ( A B ) =P ( AB )=P ( A )P(B)
P ( A|B )=
P (B ) P ( A C )=P ( AC )=P ( A )P(C)
General Multiplication Rule
P ( B C )=P ( BC ) =P ( B )P (C)
P ( A B ) =P ( AB )=P ( B| A )P ( A )
Mutually Independent events A, B, C, , Z
P ( A B ) =P ( AB )=P ( A|B )P(B)
P ( A B C Z )=P ( A ) P ( B ) P ( C ) P( Z )
If A , B ,C are mut ually independent ,
then A B C' independen t
METHODS OF ENUMERATION
Multiplication Principle
P experiments with N possible outcomes each Roll a dice and get: 1, 2, 3, 3, 3, 3
N 1N 2N p (example: DNA molecule Random Variable, Space: X = {1, 2, 3}
Discrete Random Variable and Probability
sequence) Mass Function
Permutation (Ordered Sample)
Number of ways to fill n positions with n x 1 2 3
different objects f(x) 1/6 1/6 4/6
n !=n ( n1 ) 21 ; 0 !=1 Uniform Distribution
Number of ways to fill r positions with n x 1 2 3 4 5 6
different objects f(x) 1/6 1/6 1/6 1/6 1/6 1/6
n! Hypergeometric Distribution
nPr=
( nr ) ! Pond has 50 fish, 10 fish are tagged. If my
catch consists of 7 fish selected at random and
Ordered: # of ways to select president and without replacement, and X denotes the
vice president in club which has 5 members = number of tagged fish, the probability of
5P2 = 20 (permutation) catching exactly 2 tagged fish is
Combination (Unordered Sample) N1 N2
Number of subsets of size r that can be
selected from n different objects P ( X=x )=
( x )( nx )
nCr= n =
() n!
r r ! ( nr ) !
( Nn )
10 40
Unordered: # ways to selected 2 presidents in
club which has 5 members = 5C2 = 10
P ( X=2 ) =
2 5 ( )( )
(combination) 50
(Select 5 cards from deck of 52 cards = 52C5 )
Sampling With/Without Replacement
7 ( )
w/: objects are selected and put back before Mathematical Expectation
next is taken E ( X )=u 1 f ( u1 ) +u2 f ( u2 ) ++u k f (u k )
(Number of 7 digit phone number combinations Mean:
possible)
E ( X 2 ) =u21 f ( u1 ) +u22 f ( u2 ) + +u2k f (u k )
w/o: objects are NOT put back after being
selected
(Number of 4 letter words that can be spelled E [ X ( 5X ) ]=5 E ( X )E (X 2 )
with H O P E)
Dice example:
E ( X )=1 ( 16 )+2( 16 )++ 6( 16 )= 72
Sampling In order Without
2 2 2
Summary order Variance: =E ( X )E ( X )
w/o n! n!
nPr= nCr=
= E ( X )E ( X )
2 2
replacement ( nr ) ! r ! ( nr ) ! Standard deviation:
r r
w/ nr n+r 1 E ( X )= x f ( x )
replacement ( r ) Moment: x S
E [ ( Xb )r ]=E [ X r ]= ( xb )r f (x) E ( e tX ) = etx f ( x )
Expectation: x S x S
Properties of Mathematical Expectation
Mean: E ( c )=c where c is a constant
E ( cf ( X ) ) =cE ( f ( X ) )
Let g ( b )=E [ ( Xb )2 ] , b= minimizes g (b)
Variance: Var ( aX +b ) =a2 Var ( X )
SD ( aX + b )=|a|SD ( X )
Exercise : E ( X )=7, SD ( X )=3 Y =2 X +3
POISSON DISTRIBUTION
E ( Y )=2 ( 7 ) +3=17, SD (Y )= 2 3=6
Bernoulli Distribution
A set of experiment trials with random outcome
either success or failure (tossing fair coin)
Binomial Distribution
Number of successes in n fixed Bernoulli trials
(number of girls among next 50 children born in
Chicago)
f ( x )= n p x ( 1 p )
()
x
n x
P ( X 2 )=1P ( X=0 )P(X =1)
Geometric Distribution
Probability of the number of times needed to do
something until the first successful outcome.
Number of Bernoulli trials up through 1st
success
f ( x )= p ( 1 p ) x1 , x=1,2,3
1 FACTS
Mean:
E ( X )=
p The quantity median is NOT the same as mean
Variance :Var ( X )=(1p)/ p2 Two events A and B are independent if the
conditional probability P(A|B) is same as P(A),
provided that P(B)>0
Negative Binomial
Number of trials up through kth success Event A and its complement A are mutually
x x1 p ( 1 p )
( )
r1
r xr exclusive and mutually exhaustive
Mean:k / p Variance :k (1 p)/ p2
Moment Generating Function (m.g.f)
MIDTERM 1 PRACTICE PROBLEMS A B=S
Set Probability Theory
A B A is a subset of B If A occurs then B
Notation Set-Terminology Interpretation occurs
A implies B
S Universal Set Outcome Space
Set-Theoretic A occurs, but B does not
AS Element of S Individual Outcome A difference occur
A S Subset of S Event (collection of De Morgans Law I: ( A B )' = A ' B '
outcomes)
A' Complement of A A does NOT occur
De Morgans Law II: ( A B )' =A ' B '
The opposite of A occurs
A B Union of A and B A or B occurs
Distributive Law I: ( A ( B C )=( A B)( A C)
At least one of A and B
occurs
Distributive Law II: ( A ( B C )=( A B) ( A C)
A B Intersection of A Both A and B occur
and B
A B= A and B are A and B are mutually Theorem 1: P ( A ' )=1P ( A )
disjoint sets exclusive
A and B cannot both Theorem 2: P()=0
occur
