TOPIC: DISCRETE DISTRIBUTION cell: 9360162515

I) Discrete P(x)/f(x) Mean Variance S.D MGF Cumulants Char p.g.f Moments
Distribution (μ) (𝜎2) (𝜎) Fun
1)Binomial nc x p x q n x np npq
npq (q+pet)n n log(q+pet)n (q+peit)n (q+ps)n 0  1  0
B(n,p) 2  npq
x=0,1,2,3,...n
Mean variance
p+q=1 3  npq(q  p)
>
q=1-p 4  npq.[1  3 pq(n  2)]

2) Poisson e   . x λ λ  e  (e t  1)  (e t  1) e  (e it  1) e  ( s  1) 1  0
Distribution mean = variance  2  3  
P(λ) x!
X=0,1,2,....λ>0  4    3
(K r  )
3)Negative ( x  r  1)Cr 1 p r q x rp rPQ rPQ (q-pet)-r r log.  p 
r
1  0
Binomial  p   
 2  rPQ
r
   1  qe it   p 
x  0,1,2,3,... = rq  
Distribution
r q
(or)  1  qe
r q
t
  1  qs 
f(x,r,p) 2 p  3  rPQ.(Q  P )
p p
 rc x p r (q) x Pr(1-qet)-r  4  rPQ
( pascale ) 1  3PQ(r  2)
(or)
p( x)  pq x

r
x  0,1,2....  p 
 
t 
(Geometric)  1  qe 
q
 p  1  0
 p   p 
4)Geometric qx.p q q p
Distribution p p2 p  
log  1  qe t      2  q / p 2
x  0,1,2,... 1  qe t it 
    1  qs 
0  p1
 1 qe   3  PQ[Q  P ]

q  1 p

1
5) Hyper  M  N  M  nM NM ( N  M )
Geometric   
Distribution  K  N  K  N ( N  n)
h(K,N,M,N) N N 2 ( N  1)
 
n 
K  0,1,2,3,...
min( n, M )
6)Multinomial n! p1x1 p2x2 .... pkxk npi(1-pi) Cov(xi,xj)
Distribution E(xi, xj) =-n pi pj
x1! x2 !....xk ! npi
P(x1,x2,......xK)
= n(n - 1)Pi Pj
0xi n pi p j
r
(1  pi )(1  p j )

7)Bernoulli’s  p x (1  p )  x ; x  0,1
Distribution  
(0-1) 0, otherwise 

special case of B.D

2
 TOPIC: CONTINUOUS DISTRIBUTION cell: 9360162515
II) C + S P(x)/f(x) Mean Variance S.D M.G.F Cumulats Char p.g.f Moments
Distribution (μ) (𝜎2) (𝜎) function
1)Normal
1   x 
1 2 μ 𝜎2 𝜎 
 t   t
2 2   2t 2 t 2 2 - All odd moments
N(μ,𝜎2) e 2  

 t  it  Zero
 2    e  2  2 e 2 1   3  ..  0
  x  
 2  2n  12n  3
 0
......3.1 2 n
    
 2   2 ,  4  3 4
Std.Normal 1 0 1 1 t2 t2 2 -
1 2Z2 t
 ( z) 
2
N(0,1) e e 2 2
2
2) Uniform P(x)=k, a<x<b ba (b  a) 2 ba ebt  e at log e ibt  e iat Odd moment
(or) Rectangular p.d.f 2 e  e 
bt at =0
2 12 12 t (b  a)   it (b  a ) 1  Sb  Sa 
Distribution 1   (b  a) 2
f ( x)  ,a  x  b  t (b  a)  2 
X~ U [a,b]
ba t0 b  a  log s  12
(or)X~R[a,b]
0 , otherwise
(b  a) 4
4 
F(x) = 0 , xa 80
xa
, a< x  b
ba
1 , x≥b

3
3) Cauchy
f x ( x) 
1 1 Mean - - Not -  x (t ) - -
Distribution =Median Exist
C(  ,  )
 1 (x  )2 t
=Mode n≥2
e
  x   =μ

1  λ&μ
f ( x) 
   (x  )2
2
 t
  0,   0 eit
  0,
Std Cauchy 1 1
form f ( x) 
 1  x2
 x 
4) Gamma e  x e  ax x  1 λ λ    log(1 - t), (1-it)-λ (1  log s)  1  0
Distributing ,  0 (mean= varianc) (1  t ) 
(or) Erlang) ! t 1 2  
t 1  3  2
 4  6  32
 ( a,  ) a  e  ax x  1 a  a2  (1  t a) 
! a t a 1
a>0,λ>0,0<x<∞
5) Beta Ist kind 1   Harmonic mean
1 ( , ) .x  1 (1  x) 1  1
 (  , )   (    ) (     1)
2
H
0  x 1
   1
(  ,  0)
Beta IInd kind 1 x  1   (    1) Harmonic mean
 2 ( ,  ) .
 1 (  1) 2 (  2)  1
 (  , ) (1  x)   H

(  ,  0)
0 x

4
 TOPIC: SAMPLE SPACES cell: 9360162515
Small P(x)/f(x) Mean Var S.D M.G.F Cumulants Char. p.g.f Moments
samples ian fun
ce
1) 
2
z
x
~ N(0,1)
n 2n 2n 1  2t  n 2 
n
log( 1  2t )
- 1  0
distribution  2t  1 2  2  2n
X-N (μ-𝜎2) mod e  n  2  3  8n
 x 
2

  ~  (1) k1  n
 4  48n  12n 2
2

   k 2  2n
p.d.f (  )
2
k3  8n
1
 ( x)
n2 k 4  48n
1
f ( x)  e 2
x 2

n 2
n
2
2
0≤ x <∞
It becomes normal as
(n⟶∞)(large value)

2) student ‘t’ 1 1 0  Skewness All odd moments

Distribution f ( x)  .
1      2
   ,  1  t    1 
2
1  0 1   3  ..0
 2 2     2  (or ) kurtosis n r 1.3.5...(2r  1)
3( n  2) 2r 
It becomes  t  n
2  (n  2)( n  4)....( n  2r )
normal as when r=1, it becomes std. Cauchy n2 n4
n  2r
n⟶∞ distribution ( n  2)
1
1 3n 2
f(x)= 4 
 1 t 2 (n  2)( n  4)
x.
t=
s/ n

5
 1

2
s2  ( xi  x)
n
ns  (n  1) s 2
2

3) F-Test (s ne decor’s F-distribution) Mean Variance

x 1 2
 2 22 ( 1   2  2)
F 1 11  2 2 
x 2 2
2
2  2  1  2  22 ( 2  4)
Mode Skewness
1 2
( 1  2 )
 1  2  1  2 Mean  Mode
sK  0
2
F .
f (F )  .  1 ( 2  2) 
     1   1   2 
 1 , 2  1  F  
2 2    2  2  (highly Positively skewed)

0≤F<∞

N  n  2
(ii) var ( x)    , if the sampling is without replacement.
 N 1  n
2
(iii) v ( x)  , if the sampling replacement (or) the population is infinite. (i.e., (N⟶∞)
n

6
 TOPIC: STANDARD ERROR cell: 9360162515
Def:
The standard deviation of the sampling distribution of a statistic is know s its “Standard Error” (S.E)
(i.e) Positive square root of the sampling variance is called S.E of the estimate.

S.No Statistic Standard Error

1) Sample mean : x 
n
2) Sample s.d: s
 2 2n
3) Sample variance: s2 2 2 n
4) Observed samples proportion: ‘p’ pQ n
5) Difference of two sample means: x  x 
1 2  12

 22
n1 n2
Difference of two sample s.d’s : (s1-s2)
6)
 12  22

2n1 2n2
7) Difference of two sample proportions:(p1-p2)
p1Q1 p2Q2

n1 n2
8) Sample quartiles
1.36263 
n
9) Sample medians
1.25331 
n
10) Sample correlation co efficient (  ) (1   2 ) / n 𝜌 being the population
correlation co-efficient.
11) Sample moment: μ3  3 96 n

7
12) sample moment: μ4  4 96 n
13) Sample co-efficient of variation ( )
 2 3 
1 4 ~
2n 10 2n

Utility of standard Error:

S.E. Plays a very important role in the large samples theory and forms the basis of testing of hypothesis.
If t is any statistic, then for large sample:
t  E (t )
z ~ N (0,1)
v(t )
t  E (t )
z ~ N (0,1) , for large samples
S .E (t )
Note:
Binomial (i) Mode
1. (n+1)p
 Not Integer
 Integral part→Unimodel

2. (n+1)p
 Integer, (n+1)p-1→Bimodel

3. rth factorial moment = -2npq(1+p)

Poisson:
Mode λ
 not integer →in tegralpart of λ
 Integer → λ, λ-1