(o-3)(0)

Test
proced ue foy Large somple TestsBotn)
(for small also)
compute Ho the g2uen proolem.
a. compute t
(o-s40 O-35q) 3 Specty d
Callate uale t
t- Et)
sliolas tabua
SE Ct) tiab
Sttnq of Hyo thestS
Statement com pute
differehce
Ho' - No
to Null Hypoths compare tealth tab at
HRenatte Hupothess H- rest
Tuo types of eOTS LI eol< tab oe accept Ho ot h e r o s e

’ ype 1 Ervor Rejecttn3 seject to.

Ho Lohen types ot ieste !
hcceptîng
Tuwo TaRled Test’
4y level of SRqntl cance l One Tarled Test SE eeth ey
,
type
Test level of
Too types of &amples
Sfan?Rcance
n>30
Snall ns30 Tuoo taRled testt s8

|645
’ LarQe samp Mests Rrght tatled test (>) -33

Test foy one Lett tafled test (<) --33 -|.G45

Test fox mns Sample meon
TeGt for tuo aample meay
’ Test for proportlos TSt oY Large Bample test one Lample mean
Bample propol
Sne

Srest foy uo gample (or)

propo
Vn.
 mea
for t o Samplc
test , the meon heaht of Gyoo Irdeans
’ lage Sample
56 Pnches. The nea heght of l6oo austaliafg
(o) 68. GS nches and àS nches . Do ths dota
rdans. at
nl n ndecates that oustsalans ae talles Ahan
O-o) level.

Sol
toliols
teres 64-85
) the mean
popula tion
ohose
man 68-SS
Sampe Stando deyation of a.56
5 (SoKm and Sa = asa

Lâcokm Test the slanicance

Sol:- n=uq>3O, lorqe gunple
Sarnple mecn = \5aookm 6t-85-6&Ss

pop mean (sLSO hm

ni

Los oo5 \evel tcal = -9.9 ttab =- 3 3

Ho no evert Statenent
(teall 99

Hi [Link] |teall> |tab)

e seect Ho
Austroltans Gre tollex.

*) Laxqe tample test for propottons

Test fov One

Rompls propoton
tonere

acccpt tto
 sample proportoll ) Randem Samples ot eomen and \coRty
Testtor tioo
uehey they uld PVe to hale a auctop neat
nith he gestdence. 00 men and
Uomen are SNOuT of
n the Ahe seqntfecance d2Çeren ce
blo the
proposal. Test
people n ckarnatoka tuo ropotions at S). \evel.
sheat eaterS, Giuen (TOo eample
reSt ane
Qurd the
eters
are ce Loheat
Yce
and
Can we a88ure that
t 12. level of
kanataka
State
popula en the
n=oo o (lange samde) mean Si proçorston

=0:a (cwonen prepontny

prolblem

8 colego HI PiP

(tg but Pe) ohere nL+n

’one gomple pro porton s

nn

(ouyor6)(+l
(9-)Cos)
teal 1°04
tcal = &-5 |tcall a-sa 1ttabl :96

58 )

Itoall<lttab) oe regect Ho e accept H.

e accept to. The proportRone not equa
he ce ard ohea eattis ae
to
cqucaly poputer

chaxatar,
 (l|ols

o n e S a m p l E m e a n Yandon sample of lo bays nos the folloolg 16Q (Intellqiat

) t- test test 4he
mecrn

aqneoahtey Quctient).
40, 20,
to
,88 83, to4., oo, test
4et ths ata
m e a o
aupporl the
populaon mean
aample IQ of t00.
One
freedon
Sol! n(0 -’ 80ro Snall somple test
- n i i y d:o4
Ho : 4: (0O
|n degrees
fseedom
wheves =5/ ( not qiuen,osumed s S.)
Rndependent Obseri
n-| data.
Smoll Cample test fo one 3aumple nean,
Ramole mean ß
foY to
t- test

tentn2-3)do I0o
10+120tlo +lol +88+ 83495 +9HO7
n
n2

uohere
(Io0-9:2)
(qxyco 40-q4 2)+ (i20-44-)4
nitn2- 2

39 ut5(9-8ut L63-84 4 4 4s +84-G4

nltn-2

pogre d 4- test

ten-) deo.t
Sd
t
he
2di
(teaylo62
aird)
t4ab = - a96atonm sottstics
table )
 Hcatl< ttab t y
housS n n
990
we aCCept
S6
bulbs
9S tos
lains thol
n2

Re o maufacturey

the Mean
The Test
80
houns. hours. e na 6 ( HOvse ß)
deulation of 9s
looo
bulbs
the mecn Iife of
a84 30+3a13333+9+3y4 =
&tandor
Sample not uoto gample
test)
(Small
n=Q6 &ohouTS.

L9qohouss

8).-4(34-31-98)+|aq-a8.16)4 -.a(89-88-46) J
|(a8-31-a
Ho 100O 4+6
(O
Hyi00O

q 9 0 - looo -asy

aG
tca -asy
|teall= as4 31-a8- a8-5

ttas to)n toos)a6-l

a301
- toos) as
ttab |-408 (xi-523 (yi-g)
9 1045

30 30
Me reject,Ho. 3a 3O
|63
3-6|
Tuwo horsee wete tested aCcovdRra to the sunntng capocty
fuen telo.
cTOSStrga pallculay track . The SeSuLts are
33
Hoyse A : a8 30 32 33 B3
39 3y ) S.19
Horse B : 89 30 0 a4 4 89 (9) 4-39
3y
Tes uoethev the two hoxses hae the Bane unnig coparty
Ho : uoo hoYses hoMe
does

Tioo horsec not ale the

. |teas]> lttabl
Sonple s t
' e efect Ho.
 teste
aaltolas Students
oeve Sd (3-)+ (o-G4
Q the od io
memoy Capacty

Loethey
the to\ng as Lo-l
Test
and atex tyalning follcoolra data.
Sd a48
etfectue o not fron the
-|:36.
Beore Yanig
S 2

Sol:
to : 1he tralnng CUGS not
efecve
ttab = (to2)nt
Lgs e f e t l E .

ttab= . 6
he tranig
tco) < tab e Qccept fo
5Y.
*y Testi
Sd Aectingamol
’F- tegt

n
Fo=l,n2 )

di= i - i
12 3-24
toher

5
the me taken by
2 and method qtuen below
to 33-64
Dae t h I !
Method2 :
gnant dieences blw e
Test uoether therels any
Naicneee metheod i and methOd 2.

na -l
 a0164

DR.33 6
LeneaY and Non- LRneor Models.
844334-3

na
tnulple L9near veqsecsfon
(9i - y) tRndeng he seaton betue en one dependent
Nalable an BevesGal \caYades, kon
55-05 Rrdependent
5.0
\ineay eqressen.
6 33 4006
54-45
13-s6

o.33
lety' be the dependent waxlable and trdepersdent
35 O1:80
VOTabes then the muutple inear reqression mtel
5-85
332

9= Bot BiItB22
Here Boßy 2 ase pasameters,tese parameterS are

6-1
esttmaed by luing the beloo nomal euotlons

33676

n y= Bo x t BISat B

F= Sa

Cas e

34
let y be the deperd ent vaiable ad uKL,T3 axe 3
Maniabes then the mlple luear neqresston motel i

Ftob
fere Bo BI B2ß cee paameteYS These pauamEters 3Te
F (7-1, 6-1)oros es tlnated by sollig Folloorg oma
he
= F(o5)oos
 lneoy Seqresston mo
(|.3
633
l6-| 9-3 q
323
Q96 6915
326 B2 o6
649 420 1349 I8.+
4.a
&6-8
89- depeuslent 2 aepule The MUR mo del ?s
34-

cosei,
anables

From q2uen data

Bol!
Model

he mulliple eqnession
no.o

mLR mel to the folowtno ata.

y Ft
tp2x2
NomaJ eqns y npBo+ pIZ 34 34

6& 6a 6a 6a
80
L8
.

93 8+
90 34 8+ 84 93
88 altols
the
Ho! Test rsDceduTe tov Aestug
LlG 649 30-9 18764 35a64 R06ul-646DuI pasametes in MLR
Hi! model.

l6-1 s9o34 a862 549-0) 48488. Sol6y00 Ll6-81

asus-4 5q.96 3809.8 181480l 9S-8y s$Rt SSE

94996 a6-8 a693-6 u3-89 49 32-8 87616 418-a4 (TO\al sum of aq uares)
3-4 64S 89- s8S90 &uyuy a0âg19.s 486506as 46-8I egress
33E = (sum o+ Squor es doe to
Lou3a9 349.64
(4640000 l0646 }(S- ) (Sum ot quares due to ehor)
633 3.5S34-2 Q3 - aosS 400689 los6-@S Anouo Table:
8.a Is87s a81-y|(5983 935-3 4664043 180SOl 646SG
mean sur) FcalFao
\aiGtlon fveedosn Buas of squaes
224 335-3, Regressien MSSp=$SR Fxn-k)
2 - ubGToy-3 180SOl,
SSE SSF

lotal TSS
81.9 = %(Bo) +(BiS845)+ $2 (22\4)
(Sq832 - Po(isgs) Bi(41 050|) +B(466*o-s) e acc ept to,
accept Ho means >(po)
an not tor the qen data.
seject Ho menS (poyßi Bo qiuen dat
 derence
stanRent

est the data.

nodel to the

2
Y3 2
8
134 Bo(ya)t BIC143)+ p2(26)
23 Bo . q-310

SolThe mR eqn -|138

b/uo the
- Ho: Th e e Rs no Stqn9c0te dffevence
posametess

Hi: These Re SRgne faant ditf blo the panometex

3G |6
30 SSR+SSE
36
SSE
25 12
3 3 y = 4l66,
36
25
=4-31-0(uux|-l38)
12
2 34"035 13"823 y444
30
G458
ag-36+ 5-253
36 O-69 S 0-449
5 03
2 b.695
29 333
1-359 3-34)
a-338
S u-1S 3
4-038
i0-023 02
4306 3-363 y-613 o-099
(4-355 4-202
4 691
2
1-514
Zaiy 160 0-468 Lo-023 21902 2-992
-0-Sly
 Z(yi-y) (|9-66
the prooabl |liq o occulng an elent ts qwen by P
and protobiltg ot not ocutrg an ellen t
SSP=
SSE: E(y-) 36-a03

ANOVA TABLE
e Suppose ot 4 out of to male dge QYe admitted to an
MSS Fea
Sounce of Degree of obedience chool while female doqs r
ut ot 10
VayRotion SSR
4veedon MSR= admitted en the Bame school. ind the odds vatio toy
81-3SS ' 8r-35s/2 dldmisson:
Regession F(
8olThe robabilby of male dog admtted = =0:)
n-k4 36-203
= |2-2 =y02
the prokebilty e male doq not addrnitte d = I-4 o3
n-=2 |13-66. odds male doq admsslon)e f

The prcbabi lty ot eale doa amntted

)F
he rokabllity Gemal doq ot
Feat)> FLab
oddc( Tevale doq admsston) =
.. We ef ect Ho

Regresion odds (admission) odds (Hlq aestJalue)

og*stte seqressPon the dependent \artable must k
categovlca. he logstie veqreasion mo eq atlon

-5.y.
he oddsa e of male dog adms ts
as loxqe as he emale dog admisslon.
n29
Ft the loqistle Taression
lven data Compte odd'a efo.
\ni2

nl
|SAS
 Ereplaln abot heory uesti onS
odds atio n Erplau abo ut for esom
ditterences b othem.
220 X.

NLR, LR comes, undey Supeulsed machlne leonlng

whle both Rndependent & depend ent
be tven.
equatlon

the LR model

n (s) UNIT-5

0Ruide the Aolouog data into

tao elusteYS
uging
80
k means alustereng
n22 A bfects
3
\n n
2

2
-I

tuoo cluste
dlwtded Phto
oLRVen that, the dota 2s
-(-oru2to 80G) sol

too clusteS .
lett AB, re
- (-042 tor806(o)) of too ClusteS,
’Calclte. the Otg
AB
346 2
1f c=| then ncy 5-1
2
(-0u2to-&0610)
the uore dlstones on esch obiect
= 0595
Rf x=2 then ncy)= each clustex means
dist
tormn
--0q24 0-806(2)) S*(A, Ae) - (s-t(4-)
x=3 hen niy)

oyto-068))
-688o.
oect Neest clusteY

AB

3
l3 CD.

cluster
A (5)
clustex

5431S -4-33 B=(3 2)

c=(5 2)
D-f|)

(A,ABC) = (-33-s)+(266

SBD) (-t-3)+ (-2 32

cABC) ( y . 2 ) + (QGG-)= 088S

Nearest cluste r
objcct
ABc

D
 to he
Q
rinclple

Compute Ahe
componen
penotple
components

Kolleudn 2.5

data

cooonce matrl STEP 2! Elqen AIueS ave calculated by usng

’ Calcalate the Varance

cov(*) dox tuoo ariobl

data 3.5

cov (r) y

3.5-A
6.8-4 Cad- bc =o)
S\vd det,
- (uiw) (4:u) o
n si-S) (3:5-)t 6-8-A)

(x}-5)
6-25
-A:5 st ng
Caleeted u
StEP- Elqen
3 -0.5 o5

o5.
3
Case | |
|4-5 a3y.

-304
cou (cy9)= (2)= 4y.
 2
3o||l=V()+(3-o5). 535

rom ) 535
305
G34 5- 35

s1 FPS !- Princtpal
componens

6-8-0S

3-O5 - 0 82 t0s7y

Total NaRotRon
9-84
(9s%)
Cist pinctpal com ponent)
3-05 Pe (4
L2nd) q-8utus

ariatton
eplalns
3TEfy:’ Noyal3atlon to ctn vgectors
only \atation
Pce eplolns
Vantokolee
Instead y tacing too indepeudent
Rn the quen problem Atake enly PCI
eaptan the voratlon Aue deped ent avable.
 eompou

prinetpal
ogeli uohend=2
obtaln the

62

20c-2y =0
Sol
Een values - 2 t 3 4 t 2t O
8TE PI:
224 - u t O .

3oluln g
5-À
G- 2 b
2

3 X

ohen

STER Elqen ecto S

-2
 STFPU Prlclpal compon ents

when d9.

Soluirg

-
-3

3TEPS Total Votl o n

di+ dtd
ers NoYnaslzaHon to Elae EcO
PC
d+detd3

231 3469.

lro33

2(3

- 2|3
-2]3J
 ei: epected

f\eq ueney
UNIT-3 Cot Fomla.
?nlependee of
attrtbte
test tox SON30

ei
Co-p.)
colum
(chi Squcve)
test (oi-e) obere
e;

y the folloolng table

geVeg t h e

and eye
colouY.
equeny 150

alues halr colouy

b<w ha?r =24.
colbur
assecRatton

test tE the 2s Qny

(relatlenship)
and eye olour

eye color hatr coloy ei (oi- eis (oi- e)

,Blue O-0625
20 16
S

GYey 20
20
Broisn 25 20
20
elton (associatton) blu
Ho: Thel Rs
eye colouY . 20 20
hals and
bl hoir and ee 29 24
Thel Rs assocation
coloO
12
24 o66
20 18
20 20

ei
(oi- ei)
1

3.63
ei- ous totol x column toto
GYand total 363

'.
we acc ept Ho

heve G no Telatlon blo haly and cye colbY

*9Bayes?an E ettmalon
uonen 65 hen

Posterto mean then

(-lG) +(186)

pDpalotton mean )
ayesian înenal for

aseumed Malue s Poy proabilt . G ien thal

- The
mean mark

yA proessos #eellngs abotut tolbws normal

posteviov

eQminatlon
[Link]) (64-2) )
the mean nark lles fn Ahe
Lnterva
(6s.075)
putov Probabi l4y the proteesor shol
detexmlne tue
ogslgn to the mean
mark

(4-ts
Þ, And the Osteroi maan H and
ergminatlon. S are
(Fudond de hatton) inteYVal o popuatisn mean
uo studlent yeld eng (" By eslon
conducted on nardom sample oh
teanyg and
\nte rua for
Contyuct Q5 Bayesian
Sol! tnlven 4hot

PCGS<ye o)