Let's Get Started: Unit 2 LECTURE 4: Parametric Estimation & Confidence Intervals
Let's Get Started: Unit 2 LECTURE 4: Parametric Estimation & Confidence Intervals
Hi there,
Following are my notes for the lecture. These are my handwritten detailed notes. I felt that Prof. R
taught a lot more in the videos than he wrote in the slides. I tried to make them without any
mistakes and cover most of the video material.
I however DO NOT TAKE any responsibilty if you fail your exams following my notes. So please
check the actual lecture slides and videos if you suspect something is wrong.
The notes DO NOT CONTAIN solutions to any homework problems.Only the problems and
exercises solved in the videos are included in these notes.
'If you don't want to fail, just don't do it... It's that simple silly' ~ K.J. Rowling
a
Xr * 2Xr + bXs+' "" nXr,, is also a vqndona. varualole
/^.rnrU^.r6 o
^Yt
fr:
-ffp'
a tuoe ol 6,,
iH"ilI'#'
ut^f
5L1*
^tndaxact b,{ ffi" a:; ff:^
(tn"ts ftg irnyorLar* 8g t^}s wiLL touc
f
a e_equ?y\re .^,.. or dtstribL,r.hozr und' orLo
;;v +@ cnt\,ve.rgeALe
' 3n1t^',$j;_.^ ##",
rtust kYrorro vuhA+
g9 e e
-tl^ab S, r*i.tL ataiw 6., eL,6.xnrurus+tc,valrc atn+@
--+ ry\' LwcYre'at't'g indr c'attn1
NoRt'aaLtTY
a
{svmetoTtc
tr"e w ord asymgtottc
\dluevr,euev
is used \^le "efe, *o w
6o',7{r 'n{,db
{p asttrwotorc 9n + 9is *rFt' HcaL! nsrrY't"J" tf'
)i ru, r')
(d)
-m t&^ - s) rL-+ @
or eluivaterittS E[6rl = g/
q ws see {."^4. tt^R, . Ln?r,ot thu rneo'n^ {a h}e
por,nt o}
freir,ternttst Yd*
+he eshvn,ote',t,' %
taxe m^ry *ont s o-prbs on d cqlxutdl *.tV
.irt ihe dabq irt6 6, , Oru a\€yo48 lhrs value shou[d ba fuat {o
)rh,e ,lrue, parame-tor U it +,he, uslim q+p,i. is' unfuised,.
4,4
V rcg oqkltlaho4lna biqE
Qi*t Xr, XrrXs -- - Xra, S 9zrtp) '(Vre +tueVavametex 9=P hcre
{tso \dCL exputatton is a lineqr rye*aG,f
0p^ = xn
= nrgxi
n
Xn Xr+ Xr+ Xg + - *-.Xn
l=t Yl
E [X"l = E I x,* Xr* f,3+ -'- Xn] I E [xrf+ E[xrf + --.' Etx"l = DP = p
YL rLn
E [&1 - p -- o h"nce unbaised
@P^= x (A il =
vL-
Xr*Xr
2-
E[x,l = p E Xt* Xr = I r[*,1 + -!e t Xz.1 w,*
Etxrl-p=o r-urubaised z I P
E I x,t xrl
LzJ P -O henee unbru.oed.
@ ln J1(Xr=1,X"=1)
Be$re saLvinX thig prob\e.nr" telS r.ecq,U, ufi,^t 1(A) m@v1a
n" ig the ts\d'tzqtrr varLtqbte . Tt taKe,s vaLqes o cuacl t
W,/ rtd t// # / d's/ f/,f;fiVd{drt
I+ *akeS Valu€, I r^r{tan, euen# A oearJ<,L qnd Zero AI*'U vot're'
6 -the e-xpectaUovt o[ Jl. rnre oflgt1 Pw'r.^LL^' a[ zxpzc-]akon'
Jrrtd,
E[n(A)l = !* P(4oLt;u]tL) + O;t-P(4d"e(^rhout"'rc)
ts also a$ fl, +clk4a *rH o v o$,u,s
fo E [-t(e ) ]
= P(e) rolrucl'r' ryol't' P
and coru be +l^6t/.#^t qa barwLP)
5fovo ,noYe -11 +rrkes value o 6Y I t* eac.}w ovrt<.ame of n
$L a\so {okst va\'"r'e 1 o or
9o olo es X-+
9,o dou !* or 1l
{ertee =- Il&\)z = KA)a =-.- - tfCrl
Jf(A)
po tnle C,qvu nrtf o
fn= 1' (A,) = JIt+,) Wttetg A tf {rue <-lre,n"t Xt= t n Xz =1
eonv<ar {(.)
Tf Ne h-cw a c,$nca\re,
fun.lton fC., *hen
csnczive!(-)
{ -^ zs*i"nttw :
Some {*rnulqs
> Varua-vLes
ge&"e revten: scnne
s*ort calcula*tng varriance lefs previcms sfu+t
^)e
Let *he,r^e lce Yarrclorn rza.xrabtes X,
b Y,Z
+hatore rndependent +
eoch c,they. Tneg ore nof necessorl\ idenhcal$r d^fEhib'uteel . (Bul rnaypa)
I
4 Vot(xr 1 a Vart(xr;' " "Va:t(xn) #z
-€
6L
4 = P(t-p)
Surrx,l h
A
Xl
?'
Var [xt) P(r-P)
tud
o fln= Xt+ X1
n
E[xi]= p {i
E(.
@ Dlh 1(X,=1,Xr=L)
LJE krow 111 =- ll
pn z 1l (x,--1, YI -!) = Ber (Pt)
-1nr.
luadvattc
risk { arr es{cmatorc A^ e K is
(a) r [x"l
n
-[vt"
bras & zeylc +fence x^ -l/1 is aru un]caised €:tirn"t{- t}o
=0
L z+ (i)
1
12n
4,1
tr Conflo\ ence Intcvvals
We hore, decrded o,$c stqks+dcolrnodeL 61,s (E,(E)s€O)
{nd v.re h-ow yr observakocs Xt, &,XB . "... ..Xn
NO6r hre tr"tro eluce 6- rnet^r c,arrcspf
4 *tae"ce ,i,,rierval (o,()
Yt'rst t,r.re choose cr ^urnpey d- b<,tr:eeya o to1 {e o( e
Wt-ttu the tr:F o.l aU -rdrue obse.rvadarr d.ata te vr. potnt u^e hlctt ,aofus or/\,
rrntezva,l :6 Ne ,"!ttl +alE abcrr"& ffie pra#alr/.:+1h +1,1"ah V eJ.,- N61t -thak
thr's tvtteyvaU rxLL be rnode such -thot- tt drrs" nrf depend on y ilsetf
P, tS = el ZL-d, + e€ o
- lws m adl t^ data pouT h Kt,Xe ' -' -'xv1
{*e 4uxr,a-irn,otoqq
t
n Su& th"ftt !' is colled tlr-e cnrn-fr'dun"e it''L*vvJ
ol^- feveL t- J
Wa wovut/choose d' l,be loul 'like 'o5 'Eo !-x=oQ5
jhis ev.rahLesus tp suY that -tl^e true pa'vavnel'er s h..x iw*he ti'whevvil"
t}r'aru qf 7" -t[er'r'ce -)' '{8 a conJttlence
-i.
P -u3tttL p,rdbab.t.tt{
L ErenLev
to'45. we won't g {D i'vuC wtttr, t'u?'l^ probataUr*p
'Qie
twtevval d -QeveL
S{ote, *tnohvue at^e catulahyrq Wo%diai1ifie,s . And br4 +^nd h}e nvttl *ake
tna h,etp a ddstrbuhort , r,^:fr,[ch nritl he olfe-tr-+lre -l':btbLtkrr't' * *t
d'gsgn
"?dy^fur talur 0n Nm il( 1oL \?-Le.AoYtt- tLat roe kno,nc 11,r-s
diskf buH,ovu d evevq esttw\shrr \de kvT ony 3.i *lu'ngs oLr^e
wwtch'
i
In =P" = {nt
A
tl U:l
eap 6relr p J erltnnatot [s a -$uc{c-, datn te
imnlios [G on ?tnjs b o.
& , Rr. -- - -.Rn
*
es+innat r{p c6ty1rn67t used
rrLation {ort {rso i*s a c,ornpinati"r. d taYtdar,,u
dample rnean varriobte ao i* is rtsetf o' Yqndtrntu
vaniable \^xil" a drlstributtozt
la) -xrow *o ca\aulah, pobab+Lt+ter {ot R" r^}e vnus} knovnl t.ue d,ratrttar"t+tor" o}
E, , for o. {tnite rL the distrtbwtCo'r t} R'r wtlt depend s71 horo Ri/s are
cttstributed andlindinT t+ ts o- rtXoro,r.s Pyl,c?/t9 +{or,.rewv -fox ra, +@ h}e
hone a veyg epecia\ resuJ.* ralled Cevd;rar Undt fr'eorerw f, *" ryecio&
Tondom vartr'oble SOrn'Ple wgnn 3-n
J[
E,^
#
4 , (d) > ,r) to, r)
{ n+6
E[<i]
\en 1.= = P
r-rr Rn P ,-(d)
n-+@
, ito,,;
p(t- p)
* ;; thqt
Yotcauu+1
tf*,'-' ffil
U Ip'-'l1r \, t
-lptt-p) |
-l
+Otgl
[ptt- ) p J
=L-Yt c\c.,, \s \ -dt
gL +o )rfn
=) trJe carr sotve {ort 1ct"d t'c
IP(l.Pj
)Lir4. Atso note
Jf*[l = - (r - 2"'
6-' -{) = 1e,
E(xrq.)= Q(1")
urrt"ere Q(.) isthe
1.U..is c-omes laX the de{rirnr*tot* 1*b cdf of X
(r-p)
+ )l.= \qe
fi
le> Ftr*tty w{. ve'nrrite {ch e equ o''koru
,ff,J';
*5
rj;
" l*'^hs""*$J'fi*x$t*^;ffi t
I Rn 1v,
fi
p (t- p)
, Iu + 1*b p(t -P)
JN
P (Pct4) =1-*
elEe,nnlsort I (ie p) so th'r'u 't,t vrohc'arrech
lrloLe{ilot the, La.t^exvo.l {, t+stl'p. \et( s'ee sorne obs<vvaton't
I^\ewtr\ *.. r-,.^ +o *otrru'{hjs p."tr"- but {rs{
s"t.ln*
p Bor,,re Obsewah$ns *g*dt^1 f,
O XtoG .tltot n is i+*et1 rrendovru iz it dqaend.s orn P'
this ir.^tervul wtll be cente'red artcruvtol x alLl,/\|Ju/'*
-F* e.aclrr Sonnfte
valrre 4 A"
l-- c
E'' -BE3, Ien en + 1%E)
fi_m
j.ts, winlll4 4s Rn + 9*rJptL-p
- e, o 1*, p(t-p)
= 2. lyrl@
Jn fi Jt,t
te tfs ,.,rke an tr,nfeetva)u avotrnd ffi +hat p d.betongs +o
O 1", hltdJ?r I
JN
-iii, yr1e,o6a aa vu€, im1-;reoa?' 4y-e aa' 1o'[e size +h' ru<-dtt { tP,,e ir,ntelwar(
\^IttI keel olrL decreosLn{ ie h}e r,uyitt be lrnorr- rpectfl'c ohotlti *hz
irrcter+val +i^,ar b"r#3" +" '
P
1,,vts vnzows ule a-Y^e rn -the \"t lr^t"
JtXs + eolvu the p*b)em- q fi depend.irV on P
ryts proceed
4,13
\Ne AsrD enrctier ho^c p a:( hrtll'! proloablt\ I- o( . f- t+gerf can'tain€,& p oa q.,
-tra,tevva[
Jaclox (i,.t {orrrn "t ^IpnT) 1we ear{-tdence ahouLd voV dc.pen-dcrr,p.Bo nre
rnoYe {ovn:avd to yeryroye ju'g or*o sotvre+tais prololern .
Be"{o* hle m6v€ ou, to ecr.clt o}. +tn. method it ts a g-ood r"dea to get
\zel sf,rn
+V," eon$.dence iwlervaL d " roifl, some nuynbers. fox .ttuis t.rc tirttt
+hr,
"* excmple .l:E ""1*
hL. tolt each case o{ con,ftderoce ivttevval caLcutal*on rccoll theJ.o*a,
we l"lact * = 12.4 observaUona + Ri t Bern ( p)
p toosthe *rue pavavnekv.,thot nre worttedtp esbrvrabe
Rl = * was o pvrefevrable choice s{ -th,s estry,.r,a,bor
f,Rr.
fll]ten deta r,^sqs c-ol.tecbed it wos {ound Go *t t 24 {rrrr,-sd 'tlaeiv inp(Afu rqW
W
bo ihe zsltmocrp. {or, #nJ" pot}io.tlan }ampte is En BO
+
= = . 45 or €.4.5 7.t24
Wa -lo clryab?, a ca,lFdencz trrteruval
r,'routd word
leval .e5 = l_ o(
or eluivarerrte\ a u.autd. rke\ fu 60+ ahat#at { + o( _ . 05
? ef., c€itta qs y" vr.,r*.urn
4fnt"f w=14 r ltr€ hrt[ Q*wne,-thatr tt3 ht **fl" *u maoke or ar,1^prd.z c.tr.
t}l lt^e exprzssioru thal hle calorLaLed. front *u*t ertLitx hse woul^d. r,r.eercl,
{he 1.S : ]. ozs
& qwbe.
+ q^L trt+W toh[e,E e,r *syr'e s o\.U,oave .t.rru t*rd ittu ba ,
t Rn
q v,
Jn-
JFt, _ rj ) R" + eml
_ml
Nouo .Fpl 4! a?
+ Hrp,"p) 1k + fl,*
ffi fu
1E. + AJfi'
D{vurtrotrtrl
lv,
d
Jn-
A.Jtr ln .1ln'
d
?hig ts our( ).
It? e I with pmrbAbr% qS Z " *toen p will detrr*t% Uato,^6 *r oukrt ,tnLerval
*tln ? atfeart q5'/.
*leruce {h. rne\^l c"',n
f1aence i,-te.r<vaL
arj
-Lcovrs
13
@ Rn
ery-tn
,L\5
t4, '+t^a valles
Jt= E4 = lt.tgr .ozo
1
= 1.16
n l(Eg
} C'YL
[,rre , ,lzzf
t ,5e , .721
It.l5
Solvtn4 Quadvattc erya\tow
T
,bo b6 the dey ca ol. lewt 1-D( hod
+ t,ue
r^rhere L r- 2-{
ow"\ c--*
a" l+ \*t, [= - L'1& + nl
n
shonr"t irru tha
a\? I a/\d p- *t+e, {no rort( ',osu,y
,)fme ,ov$, cort be {oun d 9ri )lrnv achttS^ y"u'lb
t'4v" cr"
-b * bL - Aac Ttru iwtarval I p, , ?rlv.illbe 41.',e
?,, I,-=
Zo-
Lrrss,{
t (n=.6q5 1'r"u- J.1( w=lt\
A = l+ U,qe)L = L,oZ
tz+
b + - L,ZZ
[*r..ns H?']
- 4L&
= l.L+SSL
C
(w" cluahm bec.omet 1,oB p2- -L,vt-7 + '4Lfi 1o pt= . SB
d P, = '7S
Wva t 'Ey r.J5 I
4.16
@o\tu
Look eoraetult/ of *he erlument below s*ep by step
Contnuotts (Y"rr
tfr p {,, ,k tlne ctrnatort *hat rnre s*qrbed
"
R'"
ff;?f *pi.+as[x]=t 1$ ft.: '(g c <nr,tinuoru
fli) qA -[o'c ftctl = fxcl]f) f<rnctiovr-
con{irusus rr1aflS^X thesrem *ella (d)
P's/F /
t-
I Rn(t-Rn)
E/a,s
n'+@
p (t-p) .
1r
JpEir {(r")
\ h+6
r {t)
Jp(t- p)
p(t-.e)
StuIsLy rs Seorern
0rr t,rl€, ean, d.r,aectty eoly- +t^t" $* I c>c) = Jprl-6
x([-x) Qirnenttnoo seluencer
P(t-P) Y la.g T T (d
y\-+@
T
n+6
G(r-il) U" -€--, ,1"
)
t* ule, ,D-lotne Jh
( E^
fl. (t-[n)
P tar
v1.-+@
-il,u,,
f,fu" t E' W,
In'
Ftr -01
1 ,
R; + 1ot,
s
$ (r -?)
1
@ I
Jplu3in f o.oa , o.Y3l