lo Bases of Statistcs And

a<ol2023
probabiity Theory
* Applicat'ons îo Statistis i
9tatistcs ?s a collecton of methods ohich belps uS
to deçcribe Summarize,, ioterpret and anadye the data.
To Researcb
)HOO medica iotervertions belps us to reduce ttbe hurden
of certaio viral disease.
2) HOco personality relates to decision making
3) betber a oed fertilzer
icreaseS-yield of crop
Raodorm Experiment 8
experment îo obicb outcome s oot certain
siogle outcome but there are foite numbers of tbe.
possibilities that may bappen
eg.Tossiogof coio
Uorandop experiment
experirnent io obich outcome is certaio.
Siogle outcome
Ha t O HoO (wgter)

Population -
Collection of all uoits îo the
expenment isclled
as the populatior.
Tt is denoted by
Note i Uoits io ao
experiment are all p0ssible outcemes
of an experiment
e areinterested io collecting inforrnation ofthe
Students_ partcipatiog fn a statistíca) Course
 Varable
1f coe bave specIHed the population of tho
Poterest for the specific research question then e
Can ibok of tborackristics of ourobservato.
Partícular fe ature of thÉs obSevatiod is called a
Variable

To otber coords. Variable is amappYng from

CSoroe set)
wbere, O is tbe Range ot mappig -

Notation We use x'to den ofe the vangble,

1alDays use capita) leters to denote the variable 7.

eg-0Cons fder the observation contaioing map of 50 pe oples

certaio ity Defne at least five variables oo
4

this populaion
} So. pe oples o a city

cobere, X, is Gender
Xa n 3 o, Io6
wbere. X2 is Ag of the person
9) } 43|6o2, 4|O06
obere, X3 is pio cod
4) }Black, grow0, $ray
where, X4 iS ee col yr.
Xs
obere X5 is No, of vehicles

Note : Merobers of are denoted by

i.e.
a. Consider. us to be a car. So, nisa.colle ctionoffcar
oefoe at least o oriable on this

Colour of the.car
-e.X (a9 ) = (oloyr obite, Çray, -

X2 Origin of-the Car

i.e: Xa (us) Origio) Ifaly, S.korea, Todia.:
* Quaotitative rariable 8

These variableg representsthe measurable quantities

Also, they can be ordered io a logical or a natural oay
i) Io a collectioo of &hoesi shoe size is voriable
To a collecion of houses, No. of rooms io
the bouse as yariáble.

Qualitative Variable

These are the variables that cannot be ordered

ioa Logical or _oaturl oay
-P9- ) Eye colour
Ty pe ofa Transport
i)
Name of political party
Remark: It s common to assign number to quantitatve
Variables but for quaitative also ûe cas assign numbers
for simplicity of analysi
eg x' be gender of person
for simplicity wwe may denote male by Oand femae
by I and vice - Versa"
 Discrete variable:
Varables obich can take fioite oumbers of
and cobch are countable even if they are
iofioite values
in
the number are called as discrete variable.
) Eye colour
sioce, is foite > Black, Gray, Brocðn ?
[discrete ond qua'tatve ]
) No. of Semester Studied
Caiscere and quantitativeJ
11) Sze of shoes -
discrete and quantitaive].
Contouous variable. :

Variable cohich cao take fofoite oumber of

values.
eg Heght of tbe person

Tf rongeset of variable is a subsel ofset of'allthe

1oegerS or cobole of Z or R;then Vaioble is discrete

TE the topge ge of varigble is iobole of Ror

Subses of R then it is contiouous rariab le, ( Real no.)
 3|1l23

Nomioal Varable -

The values of noma våriable connot beordered

Status of an applícation
f is a collectin of application J
Range set C Pendiogi Notpendiog 1.
Ordina Variable :

The values of ordioa variable can be ordered but

the diferenCe between values cannot be inter preted in
logica) way
The education System

Categonical variab'e:
Tfthe data is available io q30up form e cal the
respectit varnable capturing this. ioforrmation as a
Catgorial variable agrouped ariable
eig. statug of an application.
[yesNo kiod of ]
Every binary variable is a cateqoriml variable. hut
the converSe is Dot
Test. o0e markg of studert Ais student in
c|ass.

Binary variable -
Tt is a cagorical variable ohich coo take only
two ariables
 classmite

Irue or false staterments

QualiBaBive variable is aluonys qulitaive

True.

2)
Quantittive variable Can be both discrete and cootinuo
True
Discretf variable are always quantitatire
false
eg eiye coloux
4) Nominal variableg are aluways quaitativef discrete
True

5)
Cootinuous variabes are aluoays quantitatívR
True
6)
Coteqorical variables can be both quaitative or quan
True
)
Cateqorical variables are never Continuous
false
eq: heigbt of person ( cateqories ]
 Representation of data -
9 Bar diagram
2)
Histogra
9) Pie chart /pie diagram

) SuppOse there are io peoples wbo are standing in a supes

marleet queue are aS follows
M f Mf M M M,F M,M
obere,
M_ Stands for male a category
F staods for female '42 Cateq ory

Relative freq, uency tË = D

ohere
Absolute' frequenty
n = NO of_obs ervatior

AbsoluBe Fequey(nj) of jth categoxy

No: of observation io a particular categoy
ati

Relative frequency (j) of jth cateqoy .

 Bar diagram 2

4
3

Histograrn -
CaBsqories daBa into differeni groups.
plot bars for each categories oith the beight hË
hË = fi

wbere dí class oidth

dj = -j-L th
aod e denotes leff hand endporot of jcaBrgoy

hiloguom
3
Consider tbe follouing data of marks qained by 20 people
i0 an examina tion out of 100:
28,35,42, 90, 7,S67S, 6630,8 9,7S. 64 &I 69,
SS, 8 3, 12 68, 79, I6.

Drad histogram
HiStog ram

Absolute Relative
frequeny frequency
pistogram histogran
Cividing data into 5 cateqories
class iotervals fi hj
-2o /20-1 Yigx20
20- 4-0
a/20 3/1gx20
4-6 o 3 3/20 3/ 192
6 80 9|20 9/19x2o
4/20 4/19x2o
20

4
3

40 60
 Dete

* Pie diag ram&

= angle corie sponding to jth categox
Let ;
Oj- fix 860'

for eg

f.x860.1_x3
10
60

for eg 2
fi x360 DoX360 8

x360 = S4
20

x360 = S4

9 Xa60 l62

2
4 x 360 -
 Note1 :
nlben origbles consist of larqe numbegof differeot values
uoo bar cbart will be usefud_ but we should choase
oVerbar.
histogran
Note 2:
Bar cbart is usef wben no: of category are not too
|arge
Note g:
Piechart is used to visualize absolute ¢ Relative
frequencies of nomi nal f ordioal varigbles
Measures of Central Dispersion 8
Tendency
Arithemnatic Mean (Average)
Median
Quootiles.
MMode
Variance
Standard deviation
Arithematic Mean -
Suppase the gien data is of size 'n represeDtd by the
®n dhen ardthe matic mean of this
data is denoted by x
 Propertes -
) Sum of the denotion s of each variable about the
Xero.
aritbematic' me an is
Justißy.. n

Median -
To caleulate me dioo of the data arang e the
qiven data fo an increasaq order
Ktedian is defoed as folloos -
Notatfonto medan 05
n is odd

nis even

21 3,4, S6

2134S, 67
,n6

Median
22

4+5
2
= 45
 Medían diides cbservations ioto two
equal part al least
5o7. pa rts are greater equal median values and at
least 5o parts are lesS eq.ual median valyes.

Quantiles -

Quantiles are generali zation of ideas of median -

fuantiles partitions data into other proportions
e: 257. quantiles splits_data ioto too parts Such
that af least 25/. ot values are (ess than oY
egual to quantiles and 7 afvalues are greaker
equal the quantile
Ale denote quantiles in general by follaong notation
The dotalion is
)hece
Q lies beween o and L
i.e.. .0 < a < l

defioation of quontiles io qeneral:

if na'is not an ioteqes 4
kiSmallest joteqer qreaBer than na
nat hatL E na is an
integer
Jarmple o tempe.aBure raluesin deg rees celcius are given a follu:
ThÉs data is collected to plan a boliday io Bankok
tollowing ternperatures denoBes last years marimum knperal
trors decermber to 0th f decenbe
Q3, 24, 29, 30, 3O 2822,23190
 ce

LA

= 5

18.5

th
fod o. 25 quanles of the data.
(257)
Arrange data in an iocreasiog order
922 23,2428 29,3O,8080, 81:

25 Xlo = 2.5

na is not an inqer
(Smallest iotger greaer than 2-5)
o.25= a = 23
 Medían is 50% quantile f the data.
O.9 the quontiles are called as
deiles- ( dividing data in 10 parts).
If «XIO0 iS an intgers quontles qre called as percentle
bccause fthe data is diided ito 100 equal 'part

Tf a = O-2 o.t 0.6 ,o.8 tben quantiles :are called aS

quiotiles and data ioto 20 part.

If d iS 0-25Q.5, 0.1S then quantiles are called as

quartileS: (Making 4 parts
Mode 8

Mode Rs he value ohich occurs the mast as

Compared with other values io the data

Measures of dispersion of the data :

Varionce ofdat :
Notation
and i ís_defoed asL

Ts also knoDn as nean Sq uare deviations or sorneimes

we may called if as square mean error.CMSE).
Positive square root of the variance is called as sBandlard
deriaton and denoted by `:
Suppose there are 3 64uderhs in cass Maroj, vhas
and Mabesh. her arvas Aings in 4he clas3
are recorded in rninutes in first eo wPek Cattudoe
4andord dexfalion.
4 6
tManoj
vishal
Mahech 4 4 7

1002O
e,ast pohahility thoony ?

frponlnent

the s0o alpokalr outones in a randu, enpatnet

: otled as smple spate
) Tostinq a coit
6 ?
 4 Event i
Event îs asubset of a somple space
Ne denote it by A ,C
A CS )

obere. A Îs an event ohich is subset fa

Let ti a life in years ot an electonic bulb.

S- 0, 1, 2 9, t{
A:ife time of a bulk ith ateat too years
A=2,8,4 cs

5. Stmple £vent :
14 is an event ohich is Contaln ooly One elementof
a
sample spate.
eg Consider 's' to be Si Heart club, &pade, díamond ?

Congider
AaA Card af kinq-.of diamond
Aa is the Simple event

Compound Event
Tt is oDe that con beexpre ssed as an anlon o[ 1he
simple event
e A is the conipound event of wstat S2
Empty Space Lalul space. 00 element
Subset ofthe sample space contaYning
is calledas bull spate f empty Space -(p
micro-orqanism witbnake
eq- Consider an event of deheching-
deteching an
A o?
e2) is a set of al such Prime no of fator
thet,is non-
 Iotersecioo of ASB: A8 C S

A08

Union of A §8:

Mutually Exclusive eveots :

Ia_a certain statistica) experiments A48 cannotocCur
siroultaneously
Sucb events are called as Mutually exclusive events
Tossingq a coin o0ce and gettog Head ¢ Tail at atne.

eJ-2) for adie, qetHag an odd ever DO ata time

Complimen+ of an event A or AC

Complimentof an event w.rt. sampleSpace is set of

all elerments of S hat ae Dot n A -

Fundametal Priociple of Counting :

Theorem1an operation can be perform in ny ways and for
each of theSe a
cecond operation can be perform in na ways
then both the operatioos Cao be perform ioto nyne cways.
Tossinqq coío.toice. DË =2, ng e2 n.02 2.2= 4.
rbu many even three dígits Dumber Can we form usíng
i2, s6,g usio4 each digit only
24 numbe rs
S|2,S96gS6912, 192, 692 g62, SG2, 652, 26 Sa6 L
Y y I I 6 |26 216 62 62 SIS SG9S2, S9 2, 926 296
 The differeot arrangements are cal led as permutation.
permutaton is an arranqemen} of all or part ofset of
objed
eg How many differen tioear arrongements are possibet
-for &peaple A8,Cin a StaighB line manner ?
.2.3

The orem n distinct objects can we arrange in 'n'

different ways
e
Consider 4 lettes A,8.CD Fiad number of permutalio
of too objects taken out of 4 at ime'
No of permutation 4! E4*8. 2

Theorem- 'n distioct objects taken 'disHnct ways

Can be arranqed in n1

(n=t)]

(o-)|r]
Too locry tickets are to be chosen out of ao for 4i
and second price How many wys qre possible
ISL= Totnl no of sanmples fo S.
20| 20X19 38
(a0-2)| .18)
Th erefore,
OayS qre possible.
 classa.ce

Tbeorem !:
The numbe ofdistinct permutafions of nabjecis out of uohich
,obiechs are ot one kindne objects are of second kind aod
00and ng objects are of kth kind, then n L

Histact arrangements are possible.

Suppose write alL distinct arrangements of aLbc.

3 abc bca, cab, acb, bac, cba

Nouw et b = C =

only 3 distinct arrang enmeats are posible, -

by theorem)

Consider the corollar

The no of Days patitioning q set ith'n objects iotor

ells =

How many caays Can scientist be assigned to ane tiple.

0d tuo double, hotel rooms
I n=dMulip

is actually no-of Days of

Combiaatioo
selecting't
out of 'n' icespective of theix order. This selecions are al
AS combinatioos
rl(n-r))
Combination is achually pattiioning n objects intoon
Alote -
Theoremi
The no. of combinations O} ndisiot objects +aker
at a ime is equal to, n

frorm 4 chemists, and 3 physicist fod no of co mmifees

of 2chemists and ph'
3 that Can be form consistiog
4C.
Gien 24 y6

1)

Probabilfty The ony

times_undera
of
Expeninend that is rep eated for aby n0.only after completi
Set of cooditoo & atcome is kooDO
expenment S uch an experioent is called as ran"
hat
experiment

c s fA0g
Let A B be dhe to eveos AB then
i:e. both erent Cannot orcur siroulBaneO0sy
disjoiot eent.
 Tos sing a die

B geting odd nof

AOB=p

Mutually d'sjoiot | pafrwise 4isjoiot events

Let AAz An be 'o' possible evests of S
this set of events are mutually disjoint if
we say thaf

Complete decomposiHion ef or S
Subset ot sample space are events-J
A Aa
Am is called com pleke de
nif the usion of AA2
Am = cermpositon
and AiAA
of

Consider, S
a0 experiment of tossiog a coin.
HT
Event A $H cS (4 subses)
Ao $T? cs
Considex tthe
collecHon ofevents A A2
ALA2f is a Conpleke
decooposilisn of S

Tossing adie.
Gettiog a Nonprime fattor of7
7 = f
Geting
i+ is not
gn odd nurmbex.
formio l,31 5
decompoSitiao of as a sample space.
robability of an evtnt i
Lo) an rperimerh oe having possibe event h,
the ezperimert is repeated n ines we con (oun
how rnony 4imes eth posclble event has octlred
n(A1) noof times Ai
hos ocured in o ttials
Then the elalive fregucn is given by :
ncA)

frobabilBy of Ai in nA)

Consider an experineot ol tossirg foo dies and csett

-the no. on the upper face of tbe die. find probabi lity
o) siope evert.

Loploce defiraion of probability

proknbility of h
Tolal no of Barnple polnd:of

(Ch)
 bzlormolic defioahion
de dprcbati lity gieo hy h11 rorrogotot.
Lvery rondorn cerh hhas i4s probatitity int loed
rtes val Cog

Ariorn 2

Anoro 3

(tor disfoirh pariy

Corollaríes "

by hziorn 143

tlote'
hpplying probability to beth sidt,

by
 So:
iv)
|tocat coffee?ohat 1s IË
slowest157 pCx>25) =
od3914 o:
|- PCX725)= PCx25) > Served he
o-8s
pCx>)
PC the isleaves
PCX<
PCX<)x) x<*) the
fo length at
X=
table trips:
the of =|-P(z
o"263})) pobabìlttyoffce the
=?
24 085 =o-)s|- =
= of I-P(x<25) house
ie 0-IS 15 ioe
t O.6026 from
3 o-&sO& aboe
97 = that a
24 25-24 8:4S
8:35
+ obich heam
3& coill am
X)·04 yof)
we aod
missing Lo1e

find
g
es coffFod
:00

hal!
Success =
PCX=2) =
PC=2) Xi
aothe
No. hour
Prob-(Success)
0.0S7I
?= probability
tríp
=
of
?
O922
O-O 2 áps taking
3C 3
Co oith hat
(o-DSTI)
co-os7 succeS&
at
nea too
legst
of
x(o:3 . taost next
Va
0:0571)(1- (binomial
zordon V hour. &
429) toips.
takes
ssaie
O"00922
at
leact
 Lcte

TË he leaves the house at 8:35 am aod coff

is served at offce fom 8:4S am ufl g:0o
obat is the pobability that he coill oisslag
SoD: PCX>25) =?

pCx>25) = I-p (x<25)

= - z < 25-94
3.
= l- p z O 2 6 3 } )

o.6026

fCX725) = oO3914
iv) od the length of tíne above obich we fd do
slowest 157. of the trips
15/
PCX>) = 0-]

|- pC x<)

PCx<x) =|-o-)s -83

PCX x) o-85

’ as

t oz = 24 + 3-& X)·04

27-452
 sste

too of net 3 foips

Edthe probability that nea todsáp takes at least
half an haur 2 baun
Succest tip taking a least
P Prob-(success) - ? = 0.0sI from i)
n = g.

Xi No. of tips oith success- (binormial zondom V

8-2
3ca (o-os7U)(l- 0-0s7)
3C,oRs7)x(o:g424)
PX=2) O-Oo9222
Oncotia/

Pace

od viduad to sere at the

meodividual
e:j-) the length of time for one
cafeteig fs qq ran dom rqrioble baving exponentfal
istibyon with q mean f 4 4gos|minutes. Wbat :
-the probgbility that a person is served tess th an g
rmioytes at least 4 of next 6 days
so Given!
aitiog 4iroe to 1e the ordes
C+ime taking +1et served)
Syccess qettYog serred tess than 3 mioutes.
Meqn 4
=
ECX) 4
ECx)

pCX <3 ) =

3/4
O52T6
No ws
n=6
Serveo 1o less than mjo
Success : geting
Y: No.of dayswith syc cess
P C"ien)

I- P X<+)
I- P(X43)

>b(*, 6,0.5 )

|-O.6S6
rYt). = o-3431
taber o callspoisson
receired per hour by telephone a9swerin),
cervíce ís a TV. Xwith =6 Time in hours
Y between calls is exponential|y distibuted. fiod the
obability tha time betoeen aCalls is qreaakr than
15 oio.
X poSSoN random Variable
)= 6
y: time betoeen 2 calls (erponential r.v)
PCY> 15)

PCY>Is ) = -1S710

=0-223|
 Dote
Poce

Dístribution of Xi
Let
X nN(2)
Xí = X X2, -

X + X2 + Xa t. ---tXn
3

'otet ECX) E (XtX2 t.--- +Xn)

Yar() Var ( X+ X+--+Ån)

n

Varcaj)t Var(1a)t t+ Var (Xa)

2

n 2

Var(?)
(on clusiãn :
X 15 q norma) rv. coih mean and
ariance g?.therfort, Standard si deiakion
of is
 lssmate

(Aa)

X+ X2
tben

-efioat'on
Aoy function | Corgbinaton of raodoro yariables is caled
$ the statists.
X is q stat'sts

dis tibütoo 8-chi- square dístribytion ].

+2

(Weo rondoo arfable )

-efingtion
iodependent standard normal
Tarigble i-e.N(O,|) then their Sum ot square s
fs 2dËshhbuted andom_ varable oith
deqreeg of freedorn. ( para roete)
 -2

Dote
Pcae
Propertes of íshibution -

2-fstibytioo is not symmetric

2 d ístribyted ndorm vqriable cantqke only
8J Degrees of freed om n represets shqpe of 4Ye va lue
2-ishil
Tf degrees of freed om n increases , then the graph
istribyton ís more flat

Theorem -
Consid er, ttoo iodepeodent a nd om vqriqble which o
their syrm fs mn dt
Imp erample of dístribytion e onsiqered the
Sqmple Var(s

Sample
Sqmple rgiance cs Sarople
paramefer

fopy lotion =

popylatan varian ce - popuati

Parqmer

ECs)
populatior standard devíation
populaign variance
 318|3]23

2 ( aiu)2

Consider,

(a-xý+ - 4 2Cais)(-)
(ai-+(*-4+2(7-)(ai
(-z)+n(-u 2(F-u) (2 xi -7)
2i- X
There fore
this i'mplies
|>
(i-uf =s(i-+n(7-u
dide thnughout by c
-(
Teartanging
2 2

2
62
j

Co-1)s2 - (i-z)
 t- random varlable

gita -n deqree of feedo

by snd we caclude thaf the qndom Vow

Cnys follcas-distibuhiog with n-do
of treed sro
Cnew)

and o2= 6 ffod the probability t

eCs? >9)
so
pcs>9-) 6

ole denoea RHs area P (X> X)

)
find
794)4
.fC4&2 <? I0194 )
<49-176

r('>84) - r(X'
t- distribution:

efoatin
Let x and Ybe
and Y two iod ependeot random variables
2
tben the rato
uchtha
t-rand orn variable wbich follbwc
4-4stibution

-dfstibution is symmetrccnd its PDFis bel shogd Cura

[ie. it behareL as normal r

Je use t- ds butn if 6 is unknown

popl aHan parameler Sample paran etr

Therefore
1f 6 is unkooDo
 DatPocee

caclusiaT! 6is unknown then

(-u follous,tat
toith (n-1) dof:

find tbe p(T< 2365)

pCT<2-365) =?
of 7 1

pCT2 2-365) I- PCT2365)

o 975
 classmate

dfstribytion

-efoation: -sandom varigble

Let x_gnd Y be two iodep endent
then, thejr tatio
Cx/m) fs called asfrandom varíable

mi dof of x and n: dof of Y

eef of PDE of f-di'stribution i

Theorem
Ratio af2Sample varances
(Reduce t fronm sgmple size )
S Wtm,n-1
Anslaians Samaling d'stributians . aod E
maautacturer a car batkries quarantees that his batkrall
deviahan
last on an avg 3 years oith standarcl
batkrier have ifetime oF l.3.2-4.i a-5
-2ycars is manufchursr stll coooce h t is
wearth s
bottre s bavre st deviaim ot
normalydisibuted.
acea ASSum tht batery iFime s

uipopulaton Mean =3 yegrs

X : Sample Mean ?
n =5 C Samole size
5 poþulahon st- deviain =1 year
S Sample St. deyia tion =

I9+2-4+3-3-5 +t-2
5

=(·s-3)*+(2-4-3(3-(35-5- (ats)
5-1
O-81S

3.25E a t : S r r
 fdis
t f u - n o r m a
SangVle
oianee

tested manufaturer'sclai
otberwau, oe can
Teíectin area

57.

Xo.05
Xans - g. 488

claim that populahon meanof e

er
e.9.2) Achemical engine fom Taw mt To check -
gmlml of
batch process 5o0 batcher pach month Tf(ons
25
clain ae sampkes ento.05 gnd togshe will
-yalues lies hetoe cobat conclusion shau
Satisied with his claim.
Sls gmlml
from a sample that ba oean that distic
drawn Assume
Sacole st. de iation of 40 gms.
apprnyxinately.Normal
ot yields art
so:Given
n = 25
too

-to.os
S = 40
oX
we want to follgo distributfon about u'
are
checking claiming
Caie i) 6 is kooon then
tn
Ca3e i): 6 is unknown te

Tente. Caleulae t-saliíe 225

5l3 - 5o0
toa/ D e r a teotshel
to.os
 Toferences 9
Our Goat in this topie is to fiod qood egtimate of popu!
parameter.
populahoo parameter associated with the probabilíty
disti buion isap proximated oith the gien ample
parqmeter.

de" of an ugbia sed estimator -

Let,o be uny papulatBon pararoe<es. A Saraple

porarmeter A is Said to he an un biased estioator of S

X is ao uobigsed estimator of

Ts S an uobhia sed e stimator of s

?gustiy
ECS2)
To prove
 lus

:ECs)
#Toprove estiatorof r2
unbiased
2 is an
2

Recall;
Varcx) =

by CD-l)
S
divide throughout

S(aj-u)2

(n-)

apply E on both sideS

ECs2) = E SCxiu7
M
ECs')

ECs) /arc xi) sartx)

EMs

n
 hssmate
Date
Pece

Testingof Hypothesis..
Afn
To this procedure of testing of hypothesis. we examine
obetber a statemest about aparameter or given esearch
bypothe sis is t e ór False ie. to decide cohe ther we accept
the given hypothesis or to reject
Qoe sample problem :
Tf we use only gne type of sample to (ompare a
sample parameder wfth &ome
soroe fif; hypothecized value
oF the paramekN itis called as sample probleas
The data orqinates as ODe
Sample
from cle Lnedpopulat?
Two sample problem i

TE date orgingtes ia the form of tw0 Samples from

ttoo different populatiorns and we compare thes
parametre obtaioed from too different samples
this s called as too 6armple problem
or example,
A studyis corducted to
compare sa length of the tne.
requircd by men,and omnen to assernble a Certain product
pat experience iodicates that the distibuhon of tine for
both
OF timeI for women qrOup GPprnximatdy normal but the varianre
time fox men's gnup. Aisrandom less than variante of the
sampleof times for
meo R 14 womeniveo Stde de viation
6¢_5.g
respecthvely. Test wheihea itis Doe sample or two sample.
Dmhlem ato varance.
=

Soi- Let,
: Varíance tor men qDup
6 : Karance for womens qr0up
Hypethesis to cbeck : > 6,2

Given,pro bem is two sample problem

Null Hypotbesis Ho

Opposite of the researçh hypothegis thst we ag

dto sbeck is called as Nulk fbypothe ss.

O. AtenatiueHypothesis H
lesearch hypothesis wbich we want to check
or to ocR is Called as alternati ve hypothests
It deroed by H

Nete e: Ho is always oritten in tern s of equaiy u

inequality
TypeI 7 1ype Derrox &
Cont'usion fact Hois tue Ho is false

Ho is accepted Correct error

deation TYpeL

Ho ís rejeced Correctdcision
 classmate
Date.
Page

The Hypothesis Ho is hue. butitis jeckd

then we' say thiat yeat m¡ king ttpt I'enÝr
,

Ty pe T erroY -
The hypothesis H s false tbut itis aceepkd
then we Say thaf we

hypothesis thatH
CSuppose an allergist is oilliog to test thesome'cheese
at least3a of public is alleiraic1 to
pro duct explain how allerqist Couid make type D

Ho :< go7 are al lergj

HI 3 0 but H, >207 ate allergic
Cbecause, Hois havig
iype I error i equality
An allergist contlude ha23.o peaple are allergic
(Ho is ejehed) hut the fact isthat < so).
peaple are alerqícHere he makioq t4pe T error.
ype Terrox '
An allergist oigbt conclude that CHo is not acjected )
g0- are allergic but >ao/. ane allergíc in fact
CHo is in fa false)
 .

Tut-4
monufacurfng Fim is beng chorged tod
lorg
disoririnations in its biring protticeg.
ypotheçis is being tested if juy orareit:
) d What
aupe I error by tnding the rm is guilty?
so"' Given,
Type I ertor by Finding guily.
Ho io is not guilBy
H firm is guilBy
b) hal bypottesis is being eshed if juny Coronots
type 1 error by oding the tim is guilty
Giveo
Tyr I eror by finding guilty
Ho i firo is guilty
t, fo is not guilty

’ttest
Testog
of
(TOH)
’-testv

f- test
 1t is the probabilihy of type-I
j-e. prob (Aype-Teror) e d
pC4ypeI erfor)
prob ( Ho is true butit is rejcnd ]

Siasfance d level
T is the probability of type Ti erroT.
ie. p = prob Ctype-I error]
= P[ Ho is false, but is accepted]
p=f[Ho/n]

Lt q P are oot symoecally teatd ie miomizoin of

x and p sinultaneously is no possible in testog of the.
hypotheas
R2 The fais that if « ioc rease B decreses and vice
tersa. thertfort, one of the emrss geoeraly a is kept.
fiaed gnd Bis mfaini2ed

Def"of pouer o} the dest

pouer of test
| - P(t4pen eoT)
1-PC HolH]
= P( He/u)
 l is the probobillky o oaking
His
oecislm
tye. io fa
nesearch bypothesis H if
probabiliky o deechng Corect veeorch buo,
Ingoneral. we folloud the folloong ta seoe d
the hypothesls
sep 1 choose Ho H appropialely
Note that, H is gen erally dhe. stakmest w
to pove ¢ HoMo is alBernaHve to H R stated n
terms of oqualiy"
Sep: Choosefixed value of':
(gener altyy t s given as. g or. L't )
step3 Cbuoxe tosl statistie appropafalely crPlcal ngi
based on ' .

Crical rogion is tbe region obere, if tbe 4e statstc

lalls then. tlo is reected .i4 is also cal led as oel

Nor: ea olher dhon critica eion is Called as acepant

sjin f !he interval along the accepBante g
Cal led ay confidence
heeplane
Dgim

Confidence
ioerva
 - Nof symnehíc
bote

symmeic.

frono Compukcd tes stahisiu, ejed Ho if itis lying in

the crifical seion otherwlse, do nod jed Ho
OUY Conclu s/on.

f-dest

oith
f tes s ured ohen we haye two sample problem.

Tuoo tailed

One tailed Hi
CRigbl taled )
One tailed
Cieft dailed)

too tailed

ane taied
(Right tailed)

uo tailed
Cleft tauled)
 is the probobiliky
pothesis H
ot roaking
If 1t is
decislon
tye. ia
esearch by
he probabìliky of deecing Correct reeark
Tn qeneral. De follou the folloing shep:
thi hypothesis
shpl' choahe Ho fHa appropriakely
Nate that, H is generally dhe staement
termsof cquality
Stepa Choose fixed yalue of :
CgeneralyHt given a s or I )
Shep 3 Choose test statistc a ppopafalely 4 crHal egs
based on

Ctical regioo is tbe reg ion obere, if tbe teststatge

fals ten. to is rejected Eis also called as siei
area

Not Area other dhon hicalaim is called as accept

Hn th ioteval along the accepBante
Called Q) Confide

dgion
Cooffdence
ioterval
 4f = Not symmehie Date

symmeic Page

froro_Conputed deat stak'stic xejeet Ho if itis lyiog in

Ahe critcal seýign otherwise, do not eject Ho
Copclu sion

F-test'

S,2
f test fs used ohen e bave too sample pro blem.

Troo tailed Ho
2

One tailed H:s

CRigbt tailed )
One tailed
Cleft tailed) H

Hoo tailed
Citcaegion.

One tailed
CRight tiled)

two taled
 Date
Page

t oj Refer previous exanplk of men f obmen.

data given i
S)=6

Tel
Test the bi patbesis that '= 6 agains the aten
alh

(s-3)
$- (G-1)l+32
(5.3)2

(s-4) predft Chtical reg fon 4 il is f>s*

Critical region is ff

(s-5) As f.-132 does not lie io the criHtal glon

fail to eject Ho
 classate
Dcte
Pcce

notheroay '
Ho 6 =6
2.
(thoo tailed )
)
3) fvaue remaîos Same
( i : is one tailed two tailed )

Critieal regon:
dost haxe ftable for o.00s
ontos -So chanqe the K
take

Hence
f>foo5,

Sogs95 o13) 1

Galical
Therefore, J'32 cHtical region. Heoce, we fa |
o ejet Ho
 Z- -

"tA ondorn sarnple of loo et onded deatbin us dusng pe

T1:& yeors
year shows -that an ave oge ifes pon (s 11·&
Assume tal the populotien st. devtaHon is 2.9 year"
Doee 4 seems to ndlcale thot mean lífdine todo!
3 yrea}ur thon l0 years ?
 D>70

7l-3 70

Critca egio

1Zegs
Hence Z 65 CriHcal regioo
As 7=)022 E
-therefore, jec Ho.
 clsswste
Dat:

4-statstic is defincg as.

I:Tuo tailed Hoi=lo

t<ta
One tailed Ho: Ho t< tËK
Cleft tailed)
3. One tailed
tailed)

t
 Lcte,

o the past a machine has produce cdashers having'mean

thickness of o053jnches To deterrojne whether the mlc is
ín a proper oorking order. ASample of o cwashers is
chosen to wbich the mean thickoess is oo5 ioches is
obtained 4 std deviatian is found to be ooog
Test the hypothesis tbat the mlc is in prnper wooting
ordex usIng lexel of significane o-05 fool
Giren X o. 0s3 Mo o.o5
S=

4): Ho: = lo
|-e. M> o.05

2)
t=X- lo 3-622
o.o3/1o
4)
Critical Regco t>tK

t > 2 82| fom table

-5) As,t2:82|
Hence
i-e- 3-1622 ¬
citical'gm
the Ho
Rejet
t> ta
tto. o
t -839

Hence we Rejel Ho
X-Test - To make interenceg about
about
populaten sttde
(D-) s2
slalistic

Cca region.
Ho o = 6
Tuoo tailed 2
2

One tailed Ho = 6o
Cleft tailed)

One-tailed Ho: 6 6o 2

CRght t led) H, : 6 > 6o

 Linear Regression
l42
Sfqo test
wilcoxon signed te st
2selfby
Study
foss
0 Loear Regression'

Relationsbip ox associaton bekaeen too varable is

called as lioear Regression
As sociatoo bet?3 or more Variableg is called as
MuLtfrariate Relaton

e:j:1) Relatin bet a¡e and body weigbt coith distance Jume
( Muhivarfate
e2) Receven tme of patents depends upon many facd
that cm trol weght, Haemeglebi.,sugar lexrel.body
Hernp- ( Mulfvanate)

eg-3) Weatber depends on bumidty, temperakat, speed ot

winds, atmospheri preasure, eto. CMuttiva iate]
Consider, the ioeat relaon )

K_scalled as iotercept parom eter

B is called as slope pararmer -
oceaee
y iocreases then X also
1

it 0
X occease hen y_ decease

thematial Repregentatin bet o or pmore. y2rabl

Na
is called Model.
 cassmute
Dete
Fcce

Coosider thetollowing linear Model

Suppose,c0e bave n observation (1,9:), (4,9)....(Xato)

then each observation satHsfies the relation ,
obere
ei denoks error which rtpre sests deviaHon rom
the eqressim line of Xi, yi
Assume that, ei hae mean =0 and constant variante
e. E(e) =0

Ys called as response or deperdest variable ot on

otcome.
Xis called as co-varfate oI iodependest varfable or
TegresSOr
are called as regression cof icients os the
parametrs
Our Airo is to find a 4 B.
Metbod of least Squares
Tbis method Says that a ine can be fitted to the qiven
data set such that erorS Con be mioimized.
CY-yi)True value - pprok
Rven data se Velue

yod « p such that sum o squares o disBantes bet

 pegretnine

Scater plot

min eit
S -min (ytai
( i kyi are comtant iven 4 andF are paromeker

22(y -«-Bi) =o

yi = S«+
(T,
eq D and can be solved simutaneously to

regressicn ine y= a4fx

 Dcic,
Face

find reqre sslan which fts qiven. data.i

2

Given
Somple si2e = 4
2yi 22

j:e: 22
22 = 4e t6

}iyi= xi +Bs2
48

Solving
(I)
132 = 244 +36
192
244 t S (w)
So = 208" = 3

Hence
regression line is Y t 3x
i y
 Sziyi -(-7)Zzi

Nate the atin

by e ) )
S(«í-a) (9í-9)

Sxy
Saa

obere
Sy
S

Syy
&tÙx is caled fttd vegression 1ine or
fitted Model
when.pis negative asas increasey decreases'
and we Say that x and y having Negative elatton.
between them:
Tf f is positive as ocrease yalso in creases
and we say that elatim betoeen o f y is pasttst.
Resid eual:
Leti
Cai,yi) be obsered data point
(xi,) predícked by segreasion line
Theo
the difference beteen (yi- ) s called as
residual
 hlron9

c/essmate
Dule
Faoe

The physiotherapist advises 5of hit pakiests« All oshem had

a same knee surgey to reguloty perform a set of exerciser
he asked tbem to record time in miouky they pracice
everyday which is denoted by x and houa Lng :tates
to re cover thenm io daysobich is deokd by y
2

) 2L 24 36 64 39

65 30 60 60
n5

90
iai-) (yi-g)(xi-+)h-9
24 29 324 8
85 65 O.2 4
64 30 28-2 -31 87. 2
20 60 15.2
33 60 - 2-A
Su = |762g-305 (ai-ã)
y= 3o5
2=56G
Say=S66
y= 6)

So, Sxa =
Sy SB
Sy O.89
S

A
y-
6|-(0 489) 3s.2

= f3.78+ o 489 X
Ye 4.70 +0.4R9 X
 classmate
Date

we coo:deseribin aModel using regreim ioe

Note 1 for an intrease in l min o exercislng recouery tim

Y decreasese

Note 2 T! there is a aefference of fo rinatesf ererce

then dif io xCoveny time is 1O doys.

Tf tte paieot obo pratse 38 minuteu of exerzite

tben the model paedicts gecoVery timeo mia
interpolatio
(substate in this eg
X =33) and get
the value oY
Y= .-s37 - ).o 38 X
 cassmate
Dote
Fuce

A Physiotherapist advices 5 of his po

bad the same kDee surqery done, to iestsalaf them
a set of exercises- He regulay
asks th em to record bowpertorrn
they pracise. He theo long
they procised Cx, timesummorres
jo
the aveage Aime
1} takes tbem to regain mioutes ) and bo long
Y ime iodays). theix ronge o1 moHon
3 4|5
Estimate the inedr regresion line
24 35 64 20 33
90 30 60

Cai-z)
24 (yi-g)(zi-)(yí-g)
29
ai-)
35 - 924-8 |25.44
4
64 30
28-8 -3| 892
20 60 -15.2 -| 15-2
795-24
33 60 -2.2 23)·04
2-2
2- 176 484
-|20|
305
X= 35.2
He re
and Say l 201
There fore Sy 1201
S2

= 6| -
g7-537
(-l.038)x(2S")-g7.537
=91.537 - l. o28 X
 12]4|23
*Multfoear regres sion i

two or more deg ree L

iodependert
varrable

Aio: to fod begr elatton ohicb Connects y

Cdepedent
Varable)
e coe bave one depeodent and tuo independegt
given. y
aB and

Such tha

Givens
yi

coherei= L 2 - - n

Sum of squores ot errorS i)

=S(yí- y; )
appnxalue
ueválue
SSE
 classmate
Date
Page

Latffereniate oitk respec! tox, A4ß 'parhally

3(ssE)

This ioplies

OCSSE)

(sSE)
’(-iyi-*-i-f2i)
Tut
clussmate.
Dote
Page

Students final qrode mid- sem examinaton

qrades
and classes missed ar denoted by variable y.
respectiely The sample of 6 Students is alvon
follous. fit the mutliple Iinear rtgression equation
to the gìven date.
XLC
85 .74 76 85 87

65 50 55 65 55 70
7 5 3

437
360
S f = 24
2190O

Si. =1960
I934
|24

oimal equatins qre

2y =
497 = 6 + 360B1 24f2

3o020 860+ a1960 f t |960 f2 )

934 = 24dt (360Bi + 124B2

 -
1
isplaneregresion general gherefort,
20-64 P:
44-83 d=
1264| -8
8,80. -80 300
497 24 360 o
form Ra-24R 6
24'934 I36 24
30020 1360 2)9O 360
49:7. (24( 6.o 8
{AlB]
6
=B Ax
B2
|24 136oB
+ + 241934=
oP2 +136 21gb0P 3604 30020=
497
Pece
Dcte
 S ummary

ßs gn estimaBor of B. .e. p is an unbia sed esim

2) B is q random varigble.

Variance o

Sx

4)

5) Varck)
2

n(S2-n2)
Treralue ofyi ettnated value
c) sE ssR
6qm of suares of errors )ResTdual

dishibution with n-2 derees

parameters
and in chi squQrt ,

parameter is iaed Hen e,

 Date
Foge

SSA iis acbivssquane' distibutin oith (n-a) do

degner o Aredoro /s h-2 because patamefes
and p are substracBed friró yi. e
fxpectation o chi' clof
therefor

SSR

SSK is an unbiased es
estimator o (popuatn
vanante
 Syuat chi in and
ramekers too fnuolve itcouse be arntrced
reed{ dishiution
with
2n- chi-Squat iSq
jal SsR
Resfdual) errors) squares
of (Sum
of
Bi)s(vi-- ssEssR
= 6)
Yiof Trgyalue
ofyi
e ated
) (oitbnad n(22-na)
Varck) = 5)
2
4)
San
8)
varigble. rando q is B 2)
ased anis B. of
is e ie estimator an is B
umma S
 cnssMAte
Cate
foce

SSB tis a thiwsguane'diseibudin vaitb (n-a) dof!

is n-2 be causei too paramefus
and pf at subshacted frió yi.
fxpeckatim
thenfori
of chi-squart
S3R
ESSR) (h-2)'

SSK is an unbiaged esimaor of (egulaten

anante
 clssmate
Data
Piga

Reter'eyaoplkso phytoterepiland Hleace, fnd

un biased estimator o
caleulated
y=96f.4816 --|008X

Cyi-«-p )2
1a2 846 17ZIo4
24
65 35 8.7994
31 96
80 64
60 20 76.32 -\6"3 2
60. 33,. 6.81., -8-2|

SSR
SsR o4.9812 4q8|2

Taferencer (B
(No elatian beth a g49).
H = There is axlation bet a¢9
ToH Reqrcsian
o e. 5/

8) Decde teB stats

 clarsute

Therhort
Tis equal to

SSR

rearang (o-2)

SsR
 6o Quality Contro)

#Introd uction
Almost evers manuockríng. proce ss results fo
Some randorm varíable i9 the atoms jt produces
to be SOme \aiatior
and they are qlways joing
bet the dtorns produced qnd tois variaion is cal
Chance raria tion. Ths îs inbereotto the
qs proce
HoQcver, thert is q gnothex type of vari ol::
ohich is far froro being ioberent to the proces
lwhich is due to come assiqnable Cause.
due to assLn àble Calse tbere qe Some ad ver:
effect n the quality of toms pnduce:
later Cause Variato oay be du'e o faulty
getting macbint, pooT qually of ra m4 o
iocorect SoPtuqt ç se 4 may be glso due
byrmcn erDr
When the only vahiation_present is ue to the
chance and not by a8siq ngble cause we say that
process S in Cortrol th
The key poblem is to determioe ohether
detrmjoat
procesS iS in or out of control Tbe
of Con trol
o pro(ess wbether it is in or out
use of contro chart
grattly facilitated by the(ontrol
statistical too) s g cbqtts
Also it is deter miped by two numbers
Upper Control imit VCL
Loer Conro iot 'LCL
eratd
To employ syCha chart the data qener
-the manytactyring process divided fnto s
ghups and sbioup statisic sucb a std. devia
nd sabgnup
camputd.
 (e interences of 6 l std de vi )
4 interenrey o (mean) Tace

IAhen the subgroup sBatiste does not fall wi thiothe

we conc (ude that the procesc is ou
of Contro
led Contro chorts tor avgs 8-the xControl chars|
When the process îs io contro the successive torms
produced hae megs yre ble characteristics
that gre
fadependente-3T.V. cwith mean u and std devjgion
grading system.
(1 follost omal dístibation

Hhen the process s out o control the prod ueed

items follou! differen distri bwion atbes than
no maldistibutom
s we see whether the pro cess is out of comhol r
conol io order to stop qnd find what is
S
lorong
ely) let XXAyn be measurable cbaracterislics o
be
Su ccessive atom produced in the mf process
NieprejentS subqro date ize
b

Xnt t Xntat+Xan
 X2nt t Xant2 t . - t X 3 n

when io contro eacb X bqr hare Sqme mean

n

£CX) =
s2
Var(x)
Normal dishýbution

heo the proces is In canto x- hae astd.

(s/a)
hormal dístibution and Std nomal vgiable
cll be aloays in between -3¢ 3
P[-3<z<3] = p(z<a) -p( z<-3)
1
O.99%7 O.00|3

0 9974
ie. i the process is io Contol though o the
production th gnup we wowd certainly
Cxpect thaf
Noxi- 4)

rearranging the ttrms

- B6t< 36

LCL UCL
 Lte.

A manifachring prac esr prolucess stee) shaPrs baving dia.

that should be no mally dishibuted coith moan - 3nm
ang std deviation o Samples of
4 shats bave given the folln Sample averaqes.
Sq mple no. X Sarnple
3-02
2 2-97
3-)2 3- 14
4
5 303

Check ohether the proces S out of coto

G
U= 3 m

n 4

LCL 86 3- 3 (o1)

UCL =

3-)S

LC 285

4 Subgrvup No.