ONTT-1))

Dafirilon
Ffiune
ulaioselw
he eskimabes wtuch are Couined to
one
istence eh mare Uhan
sad tD estim
wil n ewral
H.euce o e Ls a
Consiskent 2sLimakor apnameter
to esimalors
necessi CRuteuion to dueosebetupun

with ConsisEenty Such cubeian with

Common propig a

i s based on vauiance T u ampling dstundin b

e s imatorS S Eneon as cena

o Consiskeut 2sbimabor oh
Tanod T2 are

9 e bave
C.etain pasamles
then 1 vS more bbu
4 YCT2)for all n
C11)

Than 2 for all dample aie

Concfder normal dudubutin &n sampling
Fzample

novmal populaluon NCM.)wwve p2 S khepn ahd

FAom
is wwbiased and Consistent esLimato
Sample mean
median (md) ls alo
baymmuay ample
t is Consistout as VOmd)-2io
tLLoiased and also An

= Do/2n

Runce is
VOMd) 0 as n

Consistevt 2sbímabor
 uce go find out best amen
iDsbimaba
ech
mabas
C and M) tbiuiehu is we

tn VC) =

2
f laje VCMd) = =
-51/n

as VC) <VCMA) we Conclude ample mean is

Lere ttCulet esimabr han adample adis,

anmple median

Most EfHueut Estimabor

o most eut esumabor wi

9 is

Vauiauce Vi and is Qu ottor esbimabor wit

Vauauce V. tan te uleny

E e is odahel
as
E =

Cbriously, E Cannot
Rrcord wila

Exemple!
XA,Xs) d z e 5
3audonm Sample (YuAa, X3
Wuucnopn mean
onaon FRom a omal Poualion wit

eskimatas bo estimate H.
Consider e folbcwing
XKa+X3tX4 +Xs
(1i) t2 2
0) t 2
5

2 tX2 4 x3
Iio) t3=-
 Whore s such that 3 s an ubiase estimator

Fine ne h ahd ta uubiated? dlage Fveng

ta ? t3
esimabor wuch is best among b,
peAons he

CovCXi) = 0
EC)= H Vas A =D
( i j - I , a --- n)

ECti) = E + Xat3 t Xit Xs

E(Yt X2 tx, + *s)

Fxi)

5 ECi)

ECt)=

= E * 2 +sCs)
()ECtz)

ECxa)

E X I +a)

) t M

-.but

Fta) 2
 Qn unbiased estimato
t ls not

is ulataieol stimaboy )
(3
(1i) E (ta) =.

Crs)»
&EC) t ECX) tAE

34+A =3M

VCti)= vC) vCxa)+VCX2) +VvCxu)+Vrs)

VCtI) =.o2

VCta) = VC)4 VCva)vCxa)

Vts = vc) + VCXa)}

So
4ot
 vCt) is luast t is e best esbimator e

Fxample :2 X, X2 and Ya t a tandom Sample Size 3

fom a Populalion wi mean Valu and Vauionce o.

nean
72 73 are th £stimatoxs used to esbimate
T
Vaue H Whore) 7)= Xi + *2 - *a in Ta = 2X1 +3 X3 -4 Xa

and 3 =
An +*a tXa)

0) Ave and Ta uaiased sbimators ?

wmbtased
auch trat T3 S
(0)Fnd te Value h

Lsbimabs P lonsistent
is 13 a
weta uis valuu h A
0i)

eskimator?
Gv) Lwwith is Te
best e stimaba?

daluiioo

Tandom Sample fom

a
XXaX3
is a
Since
ahd Vauane
Populaisn with mean

VarCXi) o and
ECXi)-P =

(iti =l2,: n)
. .

lovCi Ki) zo

)EC71) = E(X + -*
= E CX) +E (Y1) - FC3)

M-
 (1) FCa) =
E( 2 y) -

4 t 33)

&E CM)-A EC) 1 3E()

ECTa)
ECT)=

2 EC) tE(%)1 El3)

= M t u+

3 42u.

Wah =) (+ * 4*3) = .
T = dine s
mean is Consiskeut esimalo pppulaluzn man ,

wea auo of hage Numbess, îa is a Consiskout

esbimaloY .

2
a C1) Vaa ( ) tvaT(Y) 4 va1(13)

VaT C ) h vay(Xi) 416 varc)+9 vasC X)**C

Var (13) v a r t Y1) 4 Var ( X)1 var (Y>)) 3

VCT i's mnmu m 13 is bes Skimals

 Minimum Vaiauce Unhiased Estimators
04 dtalisc T 1 CX, *2, Kn) based on
ample d i z e n is duch t a t

) s waiased fo ?to), for all &EO ani

(11) 9nas ta dmallest vauiance a m tw class a

wbiased esbimalbos (o), the is caled e minimum

Nauiance nbiaed esimobr ils)

EoC1) = t o ) f r a l 8 6O

Var (1 VaT, () for all o EG

he em An M-v. U USs Onaue the toneRrAna PU

estirabos ethat if T and Ja are MvU astimmbrs

foo 9co), Then 7a emost duuly

Fco 1)= Eo (1) -co). fer al ocO

Varo (1) = Nas C), for a oe O9

Concider a nuo 2stimabor,7= C3.+12) which Cs aso

ub.iased ince

tCT) EcT)EC1)ita)
Vau aute isa
Voy C1)Ver (+72) dquare
Co16) ab
 V a r C ) 4Va(
a) t 2 lovCT1>9»)

Varch) t Var (J2)t vvar 11) Va7(T>)

Van CT) + VaCh) t&"vvant

VAT (T) (+2P)

1+ P
VaTC7) vaa(),) C

o-rúaut Conalatn
Pis kkaul pearsons

bebo.cen ) and Ta

LsimabDY
Van (T)2YaC))
ance is the MyO

cubakdabe VCTD Valu.in akove.

Vaag) CH P) > Vghcî)

CP) >o 1

P) We must e P=1 Jak T2 must one

neay Aulaion

Where and B au Constauts inokepenoduu y

Xn but may dupend on 9 We mau dane
 on both Side zn neat equalion
erpectalen
auing
Jaiing

E )=EC12= 0

Ec 71) BEC 1»)

BO
VaCT) = Var (1 p)

p var ( T2)

I= B=t /

tD-ehtii.eu r T e
Eaut ince Pc ) = +1 T e

00 2 must be ppiitive

etubstiube en neaT equation

Dio+1(Ta)

7
 Size
'ndaacn fom a
uch t a t foaa ample
homal populaion witn mean auance e o a
tatisbie i is the mast Hiciaa

biased.
est imat éna ' LS
tu4h
d o l u t o n

E CO)- E[«J
n+)

nt)

n
nt

iss a biaed esimab o, *)

NouD
v)=v n

CEi)
n

vC) n
2
=
(D.v
n+) (z)
V(z)
0
n nt ./
then vCu L)
 hWit, Respert to

vCz median) =
2n

VCE)/v ámedian) =

VC) 2 vC z median)

FRom @20 we Conclude that is most

ctiwnt estator

Consider a Aandom ample déze 'n faom a

noma popudalion with mean aha vautahce

u o
1est a iD Consistaut 2stimak hahd

hence ideutafy an iueut estimatos.

eution

NNCM,o)

N CM, OTnb)

CXmed) wN (H, io /an)

 V)= v(x) =0 asn>a -

PAom O P we afay that z is a Consistat

onsi.staut

eskimatv
med )aJN CH, Fo lan) and

Ez med) =M
Cx med) =Do/30

VCx med) =o sn

FRom ® oe ahy that

med ian) is a Consistaut eobimabor 4

VCa) = n vCx medtan )= Do/9

VC)
2) VZ) vCXmedan)
VCa median)

Xis an thucut
esbimator a
X
meian VC2)
vAmed)

9/
o. 6364 )
Jt he wwraed eskimalors ro)
cot andand e2 especbívahy and P-f
Cohiunl
be Conealio (o-4iut betwaen Tum, Thun

Ve.e- (1-4,C1-e) 4Ps\e,e2 t v1-2,XIt»

Let T be minímum vauamce ubiased estimalis

Aiven 9 2 are wbiased estimals

ECT) to) = EC»)

V 2
Vo C)
VoC1a)
V

Now Consider auotkor eskimabor

s-21, t HTe
eskmaor
wuh i also wwovased

r) E CT +)2)
 VCTa =
vCA) +
= . V T ) + v ( ) + 8au tov (,1

/2
e
Ne,ea
VCTB)2 v as v is min Vaiauce.

214H2

2)
2 (

V eg

e
+2f- 32H 20
Veea

) ( ) * "+3A( -120
20

(
 ) )
Dn ewral e 21 > ()

s l n quadsatic form

A>0 C>0 and

AxtBx tC 20

B4 AC £o

in e
eb
2
as

-)40

(p- - (/1-)Cr-))4o

(t-ae)- (1-e)C1-e9)
f -2 V22 T+4142 ( -e9 -et8,42)

Veka 1 +
(e14 eg 11) =o
Y
Y
 e Puahon

c = Ceneg1
-Nare
- b2 VBhac

Vass t vElea) -4C)x(etlz1))

eeg e g - (ete9 1)
Veeg N eeg - e)-e2 -)

eDt ea(ej -) -1 Ce)

-))
eR ve9-1)Ca-)
e
VC-1)Ceg-)) P Veeg'+ C)
 hcr.om 94 7, is a MvOB estimakos g flo) oE®

and 2 is any ottoy uhiased 2sbiabrs Y1o)wifh

fflenye <i Tun no uhiased lineas Conbinatucn

and Ta Can be an MVUE 1Co).

Cobinatuon

wil be wwbiose eskimabr I0o) }

FCT) L E Ci) +l2 ECT) =¢o) > l1+la =)

aine E (T) =
l®) =
ECT2)

We hane e Var CT)

Nar CTa

VarC a ) = VarC)2

(T.) + 2lle CovCT11)

NasC7)= Yar ( ) + l
Vn

v a r (1a) + 21ile (yo(i)in

V a r CT) 1

= var CT) t Ya Ch2+3l2 Jvavc71) tenc)

2
 ..02ee)

T (,2+ 2!Ja la (a 46- 1b2,

+
Vay Van

Var > Va»(71)

TCannot be MvUE

lwo niased estmator

Framplee. 4 and T2 be
o2 and CoelalhonPuha
8 0 with Vauiauce

best lw ombinaluon 2. and Dhati

s Re
ts Vandauce

estimatoo h o)
e t 9, 2 be ico uuaiased

ECT) =EC») = 9to)

et T be a lineay Combinaioo

given by1 4 i 2 / where li l2 are

lonska

ECT) = , ECT)+ X2 ET2)

=9a)Po)
is also an UBE 1() ift |hla )
 V1)= v( 11 1 )
vCT) v C To) +2l,l2 lov (1 1a)

We waut u urimuum Value

H.LMCewe we minimuum pinupla.

VT) =0 2 t 2l2 o) O P =o
dl

+ 2 P ipa =O

YCT) o2 2o p =

dubbalod ,

( - t o n)- Ja ( oi - Poro

(-tea)- Ja C -te.)
2

O-foo2 O-POD2
 g-fo, o2
o Oa- Po,02

-Po02 octh T
o102Po,2
7 given y th is Thu
Valuus q J
u'neas
embênalhon Pz
e aues ase

and ts Vauiuuce i s

to be duiuw fs
estimabor i s aaid
n
n th
leubain all the eusomaliin
it
panambber,
Te pAyameter.
Aample regauduug

is an
eskimabry
)
2.. zn)
Cx,a... .

onametes O based on dample Z 2

n en the poaulaium tuih ouusily 4(o) uch that ta|

Concuiena odishtbuhio hen s autiit

eskmako fe
ER Considos andon damples a, z 2 beleongs

Bernouiu populahon wufh paramiter 'p, 0<Pe

wiTh pmbabdli P

wuf noinalbulily9--P

n- 1C0lu2 n
p(Tk) )P CP
oistu buho0 (z.,2an)fiven
Conccnal
P(10C2 n)
=
P(z, n Z2n Zn)7 =)
-

is PCT=k)

n-
PC1-P
R)pC1- (
Dif2:4k

. i s cuhpíuu
nat odepend p .
ince i s oloeS

herem
Neymans facti sathon
lonolthoo fov a

Ssauy an augslauut
NetaSsa
Neu
atabislic es povde
'sta'bulen ko adlmit auiuu

factonsaho thuorem odu to MoYmann

4 6
talamumt - tC is clusslusut
vals
Caay a daple
Juut oliusei functuon
Can be aessed in 7 om
 , (tca) ha)
de
do Tta] olpenod s on and x Ony Jun

Tha valus
h ta) and h(x) indsp.enodeuts h o
Here, Te Lialiuvad funcluin

whereaCo.o) isa funclion ho and &

hr s a
funchon énoepenceut te paxaneber

nauiance jopetuy estimaboss

fdushiuent
V
is a autiteu esbimaos for Th

parametes and CT) isone to one huction

T and y(7) ds duticeut fos ylo)

Fzamples
e *a An be a Taolom dample foma

populalioo et pdf.
Q-)

adnaw that E T is utfueut fr .

a l a l u h e n

O-)

We kDw that .

(zí o) (o), 4(z2 .o). Can. 0)

D-) 0-)
-2

(zio) = 2. (zi,o) hz

Hence, b ocrisahon huOTem ti = T Xi is

Auyiunt estimabos h

NH.0
et , n be a rardom dample om
2

PepulathenFind autfiueut esimato5 fos and o

Let (H) -«euee , o 2024

Colulhon

buhin is
The pdf uuifovm dist'

C) 0g aic0

, stonise
a4b
et kCa,b) =
0 a>b

Ten a i = k Co i ) k(ai,o)

(
We knav that

Caio) =

(D, )«(z,0) eC0 2a)E(,0

9
kCO n ) k(Xn,o)

klo,2i) k (zí,o)

tmaxri7
D(0,14ien i2i4n

dotca], A)
IAheve a)]= klt)o) max ai

h
 . m in i
hr k

1-ence by achoi Salien iaos em

for 6
- Max xi ds duthiue LShiralar

a
Z., n be a Aanoom dampl rom

s lulion uits pd
- Cx-0)
Ca8)= e

Dbbaun auuttint dlatiskie for

d o l u l i o d

is,
The
The pd T given di'shrbahin
Cx-)
( e

We khaw hat

Caio) =t a
Cn-)
-(21-0) -(x2-a)

Cai-0)
e

n0
- e
=ei

Now Consider ue oodes dbabísbíc

 Y Y2, yn dtuate
Ths Oacrstatistic q
2 Y2 2 Yn
Aandemn auch that
lample
is
ohe rvaioo
amallest
Te pd T
D-)
dy.o)=n[-F(] (yo)

NoUD
- Ca-o)
e .d
9

Ca-)
e

-C-o)D

O(ya) =n||- FCy.)

-

Cy- 0) 7 .

-
n Cyi-a), 04y, 2
n e
0horwise
0

X2
a may
Jhus he ieihod funchon x

be weitten as,

no
 n
n (y,- co)
e
- hy
ne

e
min zi,o
h e (- n min xi)
dlabistie
enenee by Cut.eian e fvst odes
alatiskic fo
Kn) is a duhtfaut
y= min (x, >*