0% found this document useful (0 votes)
17 views36 pagesSM Unit 3
Statistical modeling
Uploaded by
rohitmags179Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
0% found this document useful (0 votes)
17 views36 pagesSM Unit 3
Statistical modeling
Uploaded by
rohitmags179Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
/ 36
UNIT-3
Es tamabion Theory
Sri ducion _
Sampling they te a Shady of srelettions ip bf un
| pepulatin ond a pompfe drawn tem Ik and
estimlton @ the population _paxamelet». Thin, the
sbakibticak snGeence iy Atvided int 2 povilo -
O Estisation trary. @ Rsting 4 Ayporrisis
eae a hone paramebeic
Trtewol 2stination
Esbrnaken Estimate. | Estimate :
> The Valse any parameter wet net be noun
—Swuvays . By -apptying seme tercrniques , WO Hy ging
Out La Sfabibtts’ de gtk value of unknown pauameten.
The techie, 10 Prewn 4h — estiniab'on. The!
Shab, used tb called estamater, The value calculatr,
suiig the Staltotie 44 called es Pnate -
Point estim ale ;
An ectimals by a singe Value 8 stabintit OF Obed.
We appuciimat. the parameter @ -2R unborn
peputerion, 4) brown a> pat estuads oy CStenaley gy
Sne parameter ,
| Titond Eshimate +
4 An tntervat eskimote ib 2 Harge g val oh
Ae astinals 4 pepulolion paxamdar
Gikouwa Qe peed eotinalon
7 lls
“() Unbsdsed ners
a) Efgiaincy
«Gly Gnwiteng,
“avy Sugeueniy,
iz
ld UNBIASED Ness
Ceh
VORA shattec 4% an snbiseed, inater “ag an
covege, Wf ends te tinal — assim value,
that are above 2h population’ prnarnden bluip
cima oy Geaguent and to the Some extent ase
ends te auume waluey below “pepullatien
betng eckimated ,
WW _Eetcrency :
# Te Megow to TR size g the 5 a be
estimate.
hn Shing b wilh, the Sinatles Yaad ang. cel
Mave mee chante oh pioduing san estat $
i the Population parameter Undey cons! eat (iy consisvEntey
X A ctatikte a a Conpiotent en F
” © ote the pepulalion
3g Bas he sample Size ineeases ik beemes
~paxamlin
Value the stalibtics
almest iceutain thab the
comes very Cae Te that & te population te
* Cneet - ;
ent estimoters ate mee awliabe ger Aasige
Samples,
GY) SurFicie wey
Se
An estimate iy Sugaint af At fakes be much
use @ te ingounalién in te sample “hak -né ether
vectimaler could exbiact- guem the sample ,
Detiiven Cunbioned )&)
An estimate. + ib sald t@ be -a _unbrastd
| estumarn ger te -potameten © 6G , E G@)=6.
Depuulion (Unbiased)
An iStimabn” +t of Q ig’ ealted cadpminbdele
Wrbiased ig Lun & (+)'= 6.
nae
Dagiation CConnistangy)
| An estimabm ty Jb
| tibimakn gor 6 ig
rane €
PL Im-e]In )¥neoNn, ee
scid do be ~a consigtent lum ,e .
(os) ae ple ol<¢ ;
a Um 3 _
: 4 n> 00 P [tm q>e] =0
©) ye - now,
Te check Conbisten cy
An estimator & cad to be . 7
E(t) >6,
vor(t) 30 ab nw,
Te +, and +, -tre to eskimaters 6% the
param, @, then the esdinater b, +» called
| mong — effrttiene 7 Hen the estimisy tr,
® von (1) é van (42) ——? ®
|
oe Ely eke K eskiinabbts
ree geus aes tit: i!
| fer the parame’: © | Then the e Stem dlow tal
| Gai to be mote, ebpreient estination gon 6 36
4b
vor (tx) & Yor (Ei), beta k, ARM Defuistion (Gufpiaing,
A estimates | @ O sb ead to be suggeii
estinalor SR ak conbuhy ald the shgoumation
about — the pata
In other werds , ag we wile
F (x; 4,0) = -@ Gedh (4%)
whe (ee) a» -a Condilionag dlsbubiction ap
4 quant and Undependent 6 6-
Then + so a buggicint —-etitin abou eeu 6,
Ext: Pueve that athe ample mon XK is an unbiased
estimater 6% the -pepulation mean A -
S
Covdi) : The Alstuloution of the Population ty
anRnewn -
le x, te Hy bee srondem Sample drawn Stem the
re
26-0 db a lobed
r n
G the -popwlation Variance o ~ and
ps timaker
| gue 2 O-%)* ay an UN RodatoecL estimdlan gon oo
| nel
Sak}
| be
Lee Gy, Mae Ky “4 Handem tomple -duawn.
| quem she -pepulalion whose mean vy Mand Vatione i,
co” te
| New
| a £ (xi) =H Ard vowiance vax (x) =o Yu
| renCx) = Gea om eG wh 2
} n
4 E(s*) =E Es)
ee e fe [@4)- ea)
[2 EDEN DES
ee [= Gen)" alse)
(Ci Ree ey
= I pfaen,
Be 2 [eo eg
\l
Pak Pree wpm iy (=). Gea)
et ee ane ore
n
Zion
Pn brosed —estimalee Bor o-®
one x [2& [2]
se (2@-9]
n 4
D> feom@
ora = Ge
Ex: Puove thak “sample man XK ib «a -consigtent
estimator & the population mean 1.
bs Gnviotent estimator
“4 E(&)=@ and wor CE) 50 a n> 09
Let %,,%, + ++ my bE a Sample -grerr the poputdtion
“ 2
whese mean A ard Voance 6”
| te F(a) =e yond Vor (mj) 2 on™ Ve. Exe
— Ex: TR) Xow Xn Oe Hardon obtewabions fem a
popaation avith 6 -nd Youudna, 2
Voiigy dhe estimallygayty = 3% ae jy sunbusted, Vana
4
Groistent gor 6.6% het.
w
‘
Gnsivtent — estimalow
5m
EG)=0, vaorCe) 5 8° -a» n- 00
fee a, xX, ys. + ty be a Pample gum the pepulation
Where mean @ dnd Vartan, o %
te E(a;)> 6 nd vou (= m7, Va,
New ,
E(t,) = eats)
= 1feG,) + €(2))
ee eo?) ' = 2
a & 46 an unbiased estimaks,
bo 6
Te check consioteng, :
© 5 Ln Fi) - bin 9
n 30 n
3 00
26 |
| © bain vou (i) = bin tog
n> 00 te} 007
sb Oo bo ib net a anmistent estimator gor o-
oo
3%o
b t= 24+
() ee
Now
y
t= 24,+ Ba,
—
lo
E(t) <6 e ries)
lo
~ rs - eG) + 26Ca)]
5 as [2.6 +30) = 58
10 :
E(4)
Van) = van —)
Zz
ca (4 von a) +9 me G3)
“ats [1074 9-7)
= 30 '
190
Te check Condintenyy
eee pes E (4) enn P/a.
@ . =o
3 re Van(t) = Lin 9%
N2@ loo
Fo tT) ay called mote -efgtcrent than
4g vox CG, )4 vou (%)
“Here, vax (xj) > Ode tyo.
Vox (*.) = var (3x, +20
y
“te (9 Gy macy)
= >
16 [ae + =>)
vox (t))=_to_ og >
16
Var (es) = Vor (esos) : iu
mare
= oe [ve Gide 9 we Ge)
EG ores
von (#2) = Jp ee
loo
Peciure
Var (ts) £ van (4) _Nayrtnonn's Factorization Theoem t-
A necessary ond suggcaint condition gou the
prgktice Ga bagficient estimator got 6 tn f(«
L{xe) - aCe) &()
where L [,e) we to Ltketinoed gundion @ 4, %y- Xp.
1 O) is,
| How, Ax) sy a conditions diebribetion ain dependerct
| Ge and ge (m6) wb 4 probability deni function
indiding a oinbite +t.
Then, & te a cufpicint estimator go 6,
Netw ;
Likelihood function 4 sqpot Fo. Téinke Probab dit,
funclion tive sample ehoountions, ,,%2- 6. ty,
pid syrnloct
L =e) = 7p pCai) $e Aisaute | andor
a Vasuabee,
tel
= NM f(a) fet contariuous
n vattinble ,
aes
Ex: Find uk a suggiant estimator in! -a Poisson
AwDubulion swith Tararneion A. K
sh 7 ’ ‘
—— The -peiven -ditbihutibn 2's
P@)= eta
EO, 2 61. 0,
al Lek a, Xe... ty be the Sample dawn gem
the above population, then te Ubelhsed, gundiisn ‘4
Ga) nee!
Sing [A]. Ae) where te 2a,
By Kayman 1s Factoizotion Theosem ,
L(A) > & [tA] AE) whore
a (eA = ee, /
Xe eal
Pal ot seme ;
2 E25, te ca Supiene ebbington ger the pepullit
Be teeter Ge Als. Sola eg Ex: Find out
2 le
F&) 6
se>
wo buffuient § estindlo, ger.6 lini
ve
The pdf io FE)= 1. e J LPO; 8 70.
Lee TX,
The abeye
1th
be
. New , the
6
the Manclem sample Quem.
Ukelifecd gunclion & poe
| Masur Likelchoed Estimation
Th grrouol fer - Handom sample), 2 Xy
The Ldelierd S4rction ub L [xre)= Fy) -
Chews can -ebtumtou 6% F Such that t [xe]
ub Mastimanm , Such -an' estimates ib called mi
Skint. Lt value toi Pron’ to be
ML 2otimatos.
The mie is « Shulion 6g
Od ge tea
oe
| dy om ge, <0
de>
Puoperticn Q MLE
_leopenticn Bo OMIE
W Mbe Albiogs © eins end consistent: «
ai) xg tn biased —eskmafon ovals, MLE we o
Unbatbed _ekaim ak, cathowie «MLE in Sluays oe mally
wanbabsed , (ily Tp suGGiienk stnbelte oxetg , at Vile 4
Wo be a unckey
(1) The Aisdidoution @ MLE ib always av,
oot vcall,
poumal .
bx: fink te mig @ A in f& = et
[aie d
os 2 oH]
x ofr =]
bee D wrk 4 7
a hb LO Let,
a >
OA A
dy <4 MLE ben y
ze
| ‘Be: Fina the MLE @ © an f(a) at oe oa.
| @>0. : ie
|
2
[Lek py an sty be AA tape gram, the alive
population
| cs , = 1) ye
Lia) -1T XY a0) 9 eee e
(«9 Tae FG emo
| NE Ene.) GR
| 0 6 oe
| éxy f
Ble
| an
bie. OLR Rhy ©
veo & guy
o>
Zu
rw
62a din é
Nb
&
e
in f
i © a
n
20 :
oa)
Z
a
bebe
x
Ea
ere
2 Ly
4
be
a
s
aynple,
giom
the.
Ex:
2 Fand
the
M
LE DNB = Soy N Poy b
Takwig- Leg on Bath rides
leg b= 2%; ego + 2 (n-x)) fogs (-0) + etow Cx
5e@
De © potty, Wh 9.
2
ee 2 £2 -f£OD p03 4 yg
ae TF ° 1-0 2 9gn PO" <0
> 0 = 2H. stra)
8 I-06
= (-0) Ea - 6 [en- ea; ]
= 24-0 24; - Onn 4642; ,
3 no = Sx;
. Th
NG 2°%
a fo= zi] a “ me
Ex? (i) Find fe MLE 6 A oh N (A, ©)
a,
aapuming OF Jb brown -
a » *
() Find the mee @ Sn N (4, 2) -atouening
Ab Renown
(i) Find the MF MLE Ww oe an w (a, -?) A ie
L | “Vo 4)
fle paf » fo = a
i ae
whee -O< KK,
co >o0
pet X, %o° x, be the sample ae
-peprtation .
Boa le ae)
wel
eS Tr I Ye Cc)
a o sam
=, ote? (ci) *
o Gn)
= 4 whe Cia)
a e
(o-2)* Can"
Taking fog on beth Sides,
alte on :
lel = ah Pre + (2 tap 2)
pies eee Grae
W Lyo DBF © meee pata, wrt |
Meee = 0 7O rt £2 (a; -A) (1)
oA do*
(1) MLE Bo when A de krewni -
>
Dig D partial wre
Fah aE Cay in 2 (Sy)
2 Gs - Haye
oO = 0 fF betel 2(%-)*
20” 2@
a ae TS Cara )* (it) Me @ A &o?”:
pi saarieeie See See ee
Gave (t)
Gove
Gxeiily
pai es \
N-2ymonn - Pearson Lemme. (NP, Lemmas)
“The Lemma provides the best souhicad suegion oe
ceching a Simple rl Aypothesi, against a Stinple
Vornalije Aypothenia .
The sk Gaba region (Ber) @ size OC foe
Jesking Ho $0 = 0. ogpinat Het @=6, so geen by
Oe bi sy when K Bo
Ls :
@ re: bl 2k
Lo
shoe Land Ly wie the LiRebihood undions under
Hand Hy Heopeciuely . The test based en the above
BCR i called Mat potvrerpal Tesk- Prog
Mek ty be Kondem Sample, fern F(%«)
oe -
clo ing Wp 070, Against H+ = 8, » Then the
°
(heGhood. Qincion a
L= Lx, o) = TF (a), 0) then
‘[oe) « bande W'S Le (2, 00]
lee fe, the oulical spin QB size OL buch that
whee K 20 le 2 constans
ox = rE = €o/u,] = Side —@
oOo
rand the power io
p= p [x6 of] = Sisae
me some then sities gion, size Wt, (24)
Then. = J boda —O ang
= Sua
@,
Now we ave te shew that
t-B 2 th lec the Jae segions @ and, be 4epined
QO =Ave snd ©, =BUc.
(5s) (<) («)
Sine ol, £0 we Rove
Suan 2 Styde
@Q ao
(ie) med ode zZ Si, ax
Auc
Suede + Stan 4 Stode + Sista
B c A ic
fice 2 fist —©
B A
See A EG we fave -
flide > K Stode
A A
Votre @ we get
fraw 2 & Slee —@
wee CA ° 4 Th the XLgon, oO , we Rave
Te, tN power Be oihical segion quale than
qpsk ty the power any other cubical sepion -
Hence @ we the Bop. Be Deive the BCR Je tesk Ho+ p=po against
hy: pep. sat Bincmal — dixbubsion — with
wp (prewr) and PP:
wporcamuatens
sks
7 “the -disbiububien gunction 3s
Pe) enon pq? } aon eee
Lek cestcc Sear be the SampSe Quer the
cabeve — -pepulakion.
- a
Le N p@)
= T coe pyr
as
= peti e(n-nr)
L= py Wc, O
ter
Te test Hh! Py=Pe -aguret 4,2 P=P, , The Ber
Acceuding Fo NP lemma iy
ex Enea, Po
pe G-P) ° Pr,
pg ee ee
ame
po Gps) ae
Dk oe
oh 3 cn +)
+)
we
Taking Sey 9” eee
ay fog (+) + £6-%) - ) > Bop
: 0 \-Po
Eu, Py T= Por
* [ow Ge) + a0 Es) | he onsale
4|
Le |
Geet" .
1): :
SE shen Ph> Pe, The Ber is zx; >K,,
whee K, ib 2 cenptant, Calateléd such that
|
p Lea; >hs[me} 20.
Goold: when Pps Pe pM Ben yg Su < Ki,
ee canes corotont chelated such. that
p[ #215] B) =o fx: Dene Ue BER to lsd fot Ae Ng genet:
oa Potibon -AUDbution uit — paranneken,
yards ,
A:
ws “The -digtnbudion qunction iy PC*) > _ ef”
axl
zo, \v0
xo,
kee Lore ve, Sy ie Sampled griom We above
20
To test Yo: A=o agate bh: A=,
TD gnd BeR by using NP Lerma. ,
L
tl > K.
Lo
ny 2%
ne
yl oth}
ee cig oe Mep e en Ee
ene y az
aia, ee
Hla). th
! ! | pF ex;
an (nt Ac ‘ '
3 ol ) A i
' 5
4@®
Takuig Cog ON beth srcdes
=n (Arde) #5 toe (Ax) > leg K
DEX Rog @) > Beg k + wn (4,-As)
eax, > fgk +” Gr.)
4m: (5) >
Gen Lubion + :
(1 shen X, mre 5
Eas >K, purhene She cons tant C, caked ate a Suen
trek p (2a; 26)/te } = 0"
Giy when AVS Woy
ZU SC, , wwhote the combinant ¢, caleulated
Such that pL 2% tok nt (56 2)
pee Gnclugion, a
When 0, 76. , Jne Ber ib Bay Ski , Where
K, ay cevtiant and Gilead Such ee
Pf ea ski /He) = of
peg 36 Sa 4, thtown , bet
4) Bird te pep
Ht Ae, where
neGei, a Baad of
Let 41, ae th. th
above population
Le aT $l)
v=!
es >, CEN -
1 Wek CHA”
€
EMG ..
Toe Bebo Me 4 A eMy -eosingt Wt Moa, by Np Lemna
the Boa i a PRA MY on TMA gt > 22 Deg
ant [H-a] > EK tna on ye
a [Noe Sey whew | Cus Snabant
Gndusion :
Gea Ay > AM, .
ty a tortterct caleulalig
Te er FH >c, phe G ®
Sach _thab P [&>e, [Ho] =k
red) Ay = Ay '
, : buch that
The BR YX
