MeASUS Of Cet a l T

Tende

Cetral temdenCt I t is the +endenen

LaLoicn Hhe VaMable valuRs ane

aloowt a centya ValuM Wid i scattere
known As owex tommonl
ReasOns wEramna
i) To aciewe o
epreentative valuo a the
whr daka set.
) 10 fai a a Cmpam son
) To tral beide
relatisn sip,
) To wo abont the
pop
v) To maue
deunsns
Citema g a 00d overase-
) I t shald be yigidba
deinid.
) It showd be expressnble in a
mathematical
wncHm the sbsen vatims.
m) It shonld be lbased n al oosex vati6ns.
v) It showld vot Loe
teted loy extrene
bbse vahons.
1 t shauld be alole toY rther algedome
maapulatiom.
(0 1t shdd ht be euked loy Sompang
luctuat5ns
Same Commg
p e g AvexaAE _

) Amthmete Mcm-
Let we hwe n valmes

1 ,: - ) An, Than t A.M hen is deuned

+2 t +

ov, = o r non re tape 2ata

For reqem dpe dalav

valutFreq
Ugrouped
AM S deuned a

Total n

closs Class
u Gisouped dalo Boundaa Mark Fre
A.M is deuned as
2 -ut22
2-3 , 2
dii
3, 2
 Assumphon
AssuPtion The
re ae umjoTml distributed
ovex the clasS, Le, all valu n a
class are
s s umed to lbe eqnal to he class mark.

A Mexits-
Tt is
igidy dgned
It i
a00d mathematical funetiom;
Based on au oles";
)hvailable 4o turther algelomc
)It is
manipilaions;
also easy to calculae

ADememts:-
U Tt is teted by extreme valus
We Cam't caleu lated A.M n case o a
frea"
ist w1 th opeM end claSses

ProPerti e g A.M
0 Sum a dewiatios o a oaservation fvm
the om is 2ero.
PIn SupposL we've n observahons 1 , , -,Kn
AM o Hhe oos"s
-

Deviations oh alout he AM
 we hawe fo
pYOve that

- ) =o
(
-

na -

n =o CProved)

Itall obs"s are eaual to aa constomt'e' hen

the A.M also eanals to e

Then

- h2e n e e(Proved)

w) 6upPOse f values oy valalsle X be

i22--2
nA.M is 2 we deune) a med vaiawle
Y a bx. The a moy is iven
a + bT.

By ey
2
 be d,
Let he Valus bb
+ b21 d2 Atb X2, - n a 4 ba,

l)N
$0 at b

2 i a t b)
i)

+bn2
oA
b (Proved
i) We hawe tw0 vamables x Y .
Valuds o X are 22, ) an
Val ues Y are
a, 32, --) dn
Den Z ax +lY
Then Z aX +LY

Pro-Let tte values oh Z be Z1,22y--yZh

Z, ajtb3; ¥1=1(1)n

2 0 t bh)
ah2+b
 ax + bY (Pro ved)
SuPOse we hawe two set o dloser vam ons. Fiysi
5et ontaans n olbs"s .
l P2 an, s and meom ok the tirst
set is
Seono et contaMns n, oloservationsX21, X22 ..
A2n, Omd mem a Hhe 6euond set is T2.
2n2
The meom o Hhe
Two sets a r e ombined.

Gombine2 set is givem b

nt2
Tota o all stos" in 1s4 seF

Similorly7 M2
Total all (,4 n) 6bs"s in he Lmbined

set i Xzi

+ 22i
. Mem ok he ombined set
+ z
 We howe k Tops
th
I a2e
2 e
,n,2 : -.,h
XmeMs
Then he meam o h tombi ned set is

n n, + +Y

orollar for 2 seks, ies bet" 7,

PraotWiHhsu lsss
}generakta, let 4 <,
Then weve to skaw
7
- 0

n4 +YM22
t2
n174+M2A2-Ym 74

n 2-74) o
Y+
 Agann

( 4) o

7 2 Proved)

valMe o
2 Medion I+ is the middle most
adaka set whem ut
is a3Ynged in ascend i
Or deundi order.

e o 1, 17, 14,12,9, 18, 13, 21, 16

Iwe YTeMA Hhe wwmbes, We 2et

, 11,12,13, 14, 16, 17, 18, 21

' The medlom is 14.

, 11,12,13, 14, 16, 17, 18

h e medom is 1t2 135
Cose-1 L Hhe
6 the (m1 1,
mediom
then th

uL t b
the vo. 6 olbs" re
ewen, ie n
Case I - I }
hem the mediom is the a.m 6 m oVnaA2m,
(m
slos".

For Hreq dype data;-

UnArUped- Med iom is he mddle mrt
vadue
50 the + o} obs" oeloo mudiom
shaua

he lesthn taPe m.fTeq uplo miom

is

Cysped
For malas- nnder the treq dist"
Cass bouwou frea Cwmreg ( Ram

Fa ta
F dtfat th
m m+1
Fn thht1ftn
N

N
 Au
-

Av41 oi
01U be 44.
the mediom Cla
The claM

F F

A Me k t I

Fx

Fk
A C
K Me utl

T wea wm thar t h HTE NA We wmlo rmlu

istri bute over tha cla Xx -
K+lttwen the
0ive comm lbe opPrOXIMAstd bast. üne between
u H The tawwe us iven albove.
We hwe PD R (sa)
Csn Snder AQPD,
hew tom pD
Agann Uwnider RE
ARPE them tom o = PE
D RE
PD PE
 KP

D AC
AB
OR-0P
o@-oP
OC oA
O-ODA

2-Fk-1
Me-Ak

bu = C Cclass width
Fx-Fy- dk

Me K C

Me -Fy-).
Medim LCB Mediom elass +

/Total req_ less thom CF upto the

Class
preceed ing mediom
Class
frea. b Mediom clas x class
wiot

Meit Dememts
u Com be calculated u It is wot c h iaidl
in
presenu o ope desined.
nd class
I t is wot I4 is ot allgoablL }0r
wcw
eheuted
in presna ther algeloric coalenlarims.
ObtiRs. w Not+ based on all oles"s.
()Mode
For data set, mDde
1s t value wic has
De- a

e ighsk frequn
S, 2,5, 1
5 , 2, 1, S 4,2,3,1,
mvde n n u it has
thi ighet freauny.
5 is the
o C o h i o u s Tomd
om
he qrouped trea diSt 4rea
For valuue with ighest
vamalol, mode LS h e

Fre
denv

class bo Un dary

foruula for made :-

A Deivaion
elass

x -
A-yH|
tA2
K+2
A C6E
n nd e r A Aob

Tw nmm imular tmmguS.

Thuy
AB AO bO
CO EO
Conhdey A BOD AE0A ANe im lr Ariomalis.
OD D
O EO

Ab OD
CE

AD
CE OD

AB K -
C

CE= k-t41 C

OD Mo
0 k 4 - Mo

Mo- Au
A- fk+
Mo-
+
AKH- MMo

ktl 4 A1 Mo

-- Mo y t

) Au-Mu t Mo-Xk

Mo

)Mo
 modal class +
Mode LCB O
/Frea. o modal olass - Fre ng the class

Preeedia moda
class
/ 2 . F r e , sy modal class - F r e . preL. moda

Aass
Frea. nLxE modoal class)

XClass Width

Ceometrie m eam
)
De For a set o n obs"s 1a, An, the
2e0metrM mem LS deHwed as_

For a arouped trea, tape dal wit K classes,

whre i-Class mark a Lclass

i-Frea. the class
K
n i Total fvea
A Propesties oh Ge0 metme meam-
0TowM o1 is equal to 0, 4hom Xq- 0

Allobs"s must be positive un oder that the

meaSure Ls beyiued.

In- In
m
o6 m AM o1 he 1n o vaMS
) G. M o the ramo o 2 vaalolks
-
Ratio o 4hir eometYMememS.

Suppsse
X
,2 -)An

.M o the rahn =
NoTe G.M is he mos Sutabla maswre

Cweragi Tahos in eowo mes it is use

oweYan PM Ca relotives.

M ValuL o1 a vamolole Chnaes dvetime

exponentiauy.

Itia vaue
Val at 1st tim PE = ar

2nd 1

h m PE T a

Mid poiwt
Valul at wid Powt O =&T

Iitial alut xFinal value

Hormowie Meam
T We owe n obs"s y A2, n , then the
hor mowM mem is
 For frea dist",

L per umt
Equa mt Onsidered HM
Equa mt ot consndered AM

P.T AM
A
G1.MH.M, For oww olos"S
Sol"- ( 72 . )"
2 sbs "
1A2
AM =
4 A2
2 M 2 HM= 2

No ( J u - J 7 ) 0
7

2,)2 2

AM M hew n= ti

Uci mú) wo aamti en and

2
 2t

t T + 1ty
(,22) (s )
t2t3 t y

So
AM GM n4 2 )
Partedimy in i s

2 , 2*,e m 2
AMGM dr r
f I at Bnt
2 md 1 ha
u ans h
e l n e i d e r

HaAr G. hn

,P-n : m (so n+m 2 ue AMamhHo,

+n tm G

A+ m s m+mv

nt
mA+m a n t m G

A m
A

naniH
GH
Combinana

alur

ea A G H

M
4
Ivimmed Meom o
Somltimes We ome oMUN 6 Sme
owtiers in a dataset. The AM is extremely snsithve
fe
eW e n a owtierS . On h
other
to he presene 61
hom t e mediom is eXtremela insens tive to +
presenw o O osteni o ho 0 oiers.To
cmprommse bet" s two measwxes, amo ther
measwre caled trimmed mm is Propdsed. Itis

ameaswre Dy Cetral temdenen whic is more

resistomwt ham AM b sed uss more o the
Ovan lable intor mai m n mediam
The trim med mputed oa
mea is
ordexmng he vamts o tte dataset Jram the
smallest to tte laraet omd hew 2uetn oa w.
VoueS 4Tom eat o he eds And finally
oweragin the rest o h
values.The trimmug
values delete
pexcentoge is the percentage oh
TOm eain en. A Ti m med mewn with a smalM
to 25). is less
drimi n percetasu S S
eHuted bowt Liers hom AM bwt is not
insmsitive as idiom.
s

Exonpl
Consn de h e dala on the so1l PH 10 smplus

757 69,7 76, 7 33,1.91,417,304, 1-1

The oTdex ed vams Ne-

7 6 , 31,783,1 85,48, 7:11,7-17, &0

3 sl, 3 53

Mem othe oos"s z 8 600t

Midiom t h e otoshs= 7 .
0 thtotal sbshs 1

. In ord e s to find 10%. tTimmed meam, we deleta

1 obs rom b o t sde h e n talke NeYO awevase

+heremann 8 oss is 7 9725