Unt-N -Disorete Prob abi lislic modeling

Markov chainiTt a Procest im cohich here aYe

Some imite no of stockei or otcome! thact can be
occupied t any given time. s toctey donot overlap
md CoveY
ier al posible octcome
Tn Mar koy procet ytem. ovel from stode
to anothex, and a Psob. u aocicctecd
cith th
tYanion foY each. oLetoone u callel traniton Prob.
A mtri cortaning tranition Probabilti u U call e
atanition ncdi.
A
Markov Procer citn to0 Stocter represetesl
graphicaly ad fellbco
state

Ex : Car Yetal Company

be company has wo oicel One ct
Bongalore
he ther at Tümkur. The co npany amanget tour'st
activitieu in both +Hhe Citie, Atraveller s start from
any of +he ci4iel. From he hiutory, follooing informaon
is avalable

End 4he long-term behavio r od caY rettal Coonpauny.

Sol Auplion Totel oo of caYy i constamt oluring

let Bn percentage ot caY avaulable in B lore,

Cmcl of n' Periou
Tn
T
The tranAion ctri

TO3 0-7

The ciscret e olynami cal

Bn+ = 0-G Bn to-3 Ty
Tn+) = 0-4 Bn +07n.
lel Bo =l) To =o

9-5a
3 O556

0433
5
O57I019

O.5¬13 o4

Bn=0- 4a8
In the
i.e Pproximately 431- of the cars conl be i B10xe
Tam kyY, Tt tate approxiatel4
+0 each hy statte
 Voting tendencie
Cohidey a
hrec Party Stem itn repuslicant,
DemoCxactel and independet. what s the long
behaviOuY of vote in he election2 it here aYe
no
adoitiom or dele 4ion of
votexu. The 4rend L les Stricly
dong Party linel. "The hy pothetical. hutorical data.
tor Noing trend Prejemted in the Folooin
tranitiom next sete
R
R os o-05 e"a
Presemt
state

Soln let Rn, Dn, In: be he

Percemage of votera in
Pevio d n
The dyno wnical sytem u
Kn = 075 Rnto-a Dn tO-
sTn

Tn = 0aRnt o- a D o4In
ko = Do = To =V4
Rn In O-959

O"555I 0145a

O-a833 O2668
In long rum,
O500 0-2534
o5aGI O. a50
561 q voters ave wh R
-11
0.5344H 2163
854: 41
o546
o.550
055aq
 Idea o a Marho chaini
A
Mar Kov chain ua PocH Conlisting of a
sequemce of events uoitn tthe fllocoimg
9 An evet has c propertiea.
finite mo of otconel
stete. The p called
Procek s cl0aye in Ome of these stoel.
At

outcome
stage
each
or perioe of the proces,a particular
Can tranition fro it prelet Stecte to
ny other Stcte oY e e n in
the Sane Stcte.
6) The
Prob. of going fxon eme state t anotthey in
a Singc stage u vepresemted oy a tranidion mdx
foy ohich +he etries in eoch o) aYe
between o ard.
lach Yow Sum to l. Tbele
On the present State and
pro bablities oepemd onlq
mot on. pat Steter.

Cx-6) Dining behaviouy of student

Tt u found tht 5. of the studemt ho e t
at hottel A eturn 40 eat here Ohere a
4hose oho eat at
getusn gate. dining hall haue e 93/
There only
asd

halla
akume 4hat oll +he
etuclern
diing
hally voi lable
eot ct oe of thee
Fos mulate a mel
+o Solve Jor the
Pes centage oj student
eating d each Aerm
dining hall.
 The tyanition mdrix u
hext stade

psesent a6-a5 o-7s

State

et An - be the . of stucdentl coho eak in. hall A

The diScy ete dynami cad Sytern ek the emd of

(nH) Perio d l
Ann = 0-a5 An t O0 Bn
Bnl = o75An t 043 Bn
lel Ao =l, Bo =0
The Soly

An Bn
-

O-75
O.885

3 Oq09

G-914
Qq15
6085 0-915

tne lomg 85. of 4he Sludert Choose

hall A arnd al.5 . choosc hall R
Modelimg conponent amd Stem veliabi lity-
Reliabilityi- The veliability of a Com ponent or a
bYttem thne psob. thc it oill mot foul vex a
Specitied tne peviod.
Let f)
be the foulure yate o am
item, Co?on ent
Sytem ove time t, So hd fc u a prob.
over
Let F be the Cumuletie distibdon,
oistibuctiom Sunction
Coryepomding to fc) Here FH u Cumul ative falure
te c ine
Teli ability at any timG tu
RCE 1- F
i.e the veliabil ity ect
any tinme t iuu l minu 4he
expected Cumul ative foilure
gete et tine t.
Human - machine Sytem
electonic os nechaníca
Conistu of canponent, Fer
Sueh a eq. Conlider a Sytern
your conpute, Nour stereo
your
autonobile. So on
Heye, we oat to bild a Simple
modele to
examin thne yelia bility of conplex
Sytem
Now
Conider dleli g Yel ation ship in Seriee,
Parallel oT Combination of thele.
 the
Compomentl of a Sytem Ye Connccted in
) Seriel.

The Sexie Syten i one that perfoyn

whem that well aonly
ell tne
Component are fully fundional,
Then +he
Syttem Rcli ability y
) Parallel
The parallel sytem u on e 4het pertorm
well as long al One of the Con poet functiomal.
-4hen 4he
Syetem eliability iu

Combinotiomi

R - (2,4) (eatR, - RgRy) (e)

Eg 0 Rocket
populion Suten for a space
with boostex Yocketl in shutle
Series is shou0 m belou
Rg=08
Find 4he veli abi lity of 4he Sytonn
3ab Reliability of a sytem in Seria a
R)= R Ra R = (o-) Co-8) (o.a)= 0.655a
Note thad, he vesutont veliability
of +he Sytem
ler than 4hat ot he inoivilua
Compon ent.
Eg:- Coniider the following
Find tthe reliability ot
commun
ComnuaicatoN Sytern.
the Sytem
R=0-45
migawave yatto

Fm Racio

Soln Since 4he Syten uin

of the Syt en u parallel, t+he reliability
R4) = R +Ra-RiR =([Link]) +(o-q) - (o-q5) (o.-)

Note that Syten Yeliabilty u much

greccter than thc
indivicual Conponent
eg:8 coicler the Combined Sitenn Shown below

Find the veliability of the Sytem

- (o.98s) (o-q) (0-64) (o-q)

(4 Goider a slereo Syitem Oith the Conpon en kt
CD player, Fm vadio, speake a Pouoer am plitieY ay
Shown in dig. wth
o eliabilitie. Deternine 4he Sytem
veli ability.
speaker
PA
45 UF speaker
Sub-Syton -)
S
Sa
Se RCs) = o-98 to-q) -(0-qg)
(o.q) =0
= 0-q994
R(Sa) = G-q9 t O.98 -([Link])
(0.q) = Oq998
The Sytem Reliability u
R= 6 45* R(S1) * RSa)
(-qs) Co-qqq4) (0.q498) =0.q49a.
Find the. veliability
of the Computer Sytern
Shoon beley
0-q49 o.q95
key board Pondrive
4Printer.
Poler 0.496
Supply Unit Havd disk
nouse
Leo-Ram LPsimter-a
o-493
RCS) = o.q9q t o .q98 - (o.q99)(0-q98) = o- 44949%
R(S) 0-495oqaG +0-q9 -fo.qa5xO .qq6 + o-q96x0-49

t (oq45) (o. q94) (o-q4)

R(Sa) = -qq1+ O-qq3 - (o.9q1) ([Link])

= O. qq4937
R= R RaxR(s) (R(Sa) ×
R(S3)
(0-494) (o. q91) (o.499) (o.q9)
(o-9q9)
O-495q35
 And the veliability of the Byten
HRadlio
Power q45
Gupply stovage
LiRadio a
O.q45
B.
Pa
Sa
6-44.
Soly
RCS= q45 t o q95 - (0:95 X0.q95)

R(s4) B+BatB, tBy -B, B¡t Ba B+6, But BuP

B, Ba Bg By
| 00 00|4G

R= PRs) x RCSa) x Ra x R(S)

Linear Regveioni
In thne methoo of least Sqsave, oe haue
Conhidexed anly a Single observetion y: for each

value ot the imdepem dlent Nariable Xi Naw we Study

a gtatiti cal methodaloqy for
ot the mimimniginghe
SquaYed deviadtionl, called
calledl lineay Yeqy eion.
objectivei
1) To itlstyete the basic linear Yeqrekion mdel
a) To define amd imter pyit the Stadistic
3) To dyaw the ScatteY plot of reidue
LineaY reqrion model
The basic linear Yegrcsion madel u ot the
foy-n
Cm poirta)
The mormal
eqnt are
ma tbZË=2i

[Link] stati sli cal analy, define he

folleoing
EyYor Sum ot Squqre ( SSE)

cohich ye<le ct vaion abot the Yegrkim line.

a. Total coTYe oted Sum af Suare (SsT)
SsT

where y u the aveYaqe of the y value for the

data paintl Cxi, y;) Computed from the model
3. Reqreyiam Sum of
at Squarel (ssR)
SsR= SsT - SS
The quamtity SsR repreentl the anownt
of Nari ttiom in +he Nalue Computed frem
he Yeqreion lime when Compa ed
+he Nayi oto in the y aluer abot tthe
Iine y=y
H Co- elici ent ot determimtion CRJ
R? SSE ohich u a meare of fit
SST for the regriom line.

ex pre he Psyo portiom of +he total variatio

im the Nariable y of the actual data (ohen
Co om par ed with the lime y = ) t h t Cam be
accouted fo the
Naluea are
sAraight line moel, eohose
givem oy a +bx (the Predicteo valuee)
amd Calcul ated imtey of the
 Eg- T4 R'= o-q3, then q3/. of the total Nariation
of the valuer (fvon the line y ) u accoumtec
7 by a liner Yeliom ship coitn the valueg of

u claSey +o , beteY he fit ofhe

Tegreion line.
Notei
O The value of docnot dep ernd om cohich of
Vaiablet is labelled as
The Nalues of R u imalepen det of the cnit
of

Anothex indictor of the vealomqbleme of

+he ft u a plot of esiduali Nerse he
imdepemdent ari able. Here residuad u mathmq bt
error bw th e actua and Preoi cted Valuee.

=9- (a;+b)
In 4heplot of vesi due vs
O The esidue should be reat
indepernert vaviade
amd yandomly distribted
Continued mn a
G An veasonably Small bamd.
extremely larqe ve,i
reoi dual
duau eguirea further
imvetiqation
Tremd in the yesiuall
4he model if
indicadtey ho) 4o yeine
Yefimenemt u needled.
Pokible patter of residual plots.
R

Fanning Linear
Cuvilineqr

Gg- 0 Ponderosa Pinel

Lumbey Cutter wish to use readily anailable
measur ememt o
extinete the mumbey of 'boad
feet of lunbey im a 4ree. Axumt they rneasure
4ne diameter of the +ree m
imchel at oaist
height Develop a model nat
as a fun of diametey.
predict board feet
Lumbex : CollecAive teyn tor haveted hat ha
been
cwood
manuactured ito board.
Bo axd feel: A unt ot measuYe foy the volume of
Aumbey im
1 Board feet = 1 tx1 t xl inch.

Problern idemtificodion.i redict +he n of board

feet as a fun of -the diameter of the ponderOSa
Pine.
 Akumptionli- ponde voso, pine aye geonetri cally
Similar
and hawe tne shape of ight Crculay
cylinder. alkune diameter as the chaocleristic
oimenion.
madel Sormulecti on
By geometric Similàrits
VadÔ, ohere d u ne diarnete

It we aume hat the treel hawe corkt hight

them

+hot +here Coyt Nolue

akoci ccted oi4h the tudergound "oot Sytom, hen
We vefine 4he md a) a folloo
N =

esti cte the Parameter fron the followig dleta.

0i ameter Cd)9 731 ao a5 939 33 7 39
Board feet(v)i as 16 |14 328G

Nolei NOrn al fo
= at+ bx =a+b
natb=Sy a=y
na +bE=9
3
|as +bz= y
 3
Vg = bid'by= yEdv 4aH6550g

I18878l -0: 154

774633

V = o15y d

loat31535b= lo a9
a31535a +9.61xlo'b = 4ah65so 9 Salvin9
3
b= H32 x1o

NI= a. T4a5 + o00h 3ad

7Hn a t14 633b =

I188781
a = -3 9-70 b= 0-|9

Ng =-34.10 + o-lq d
d
NH
14 3o.18 -a.18 55-59 - an.59 3a-3] -H-37 [Link]-689
17
-5.6 H4.51 - ag51 a3-91 -7:47 15-al
31 079
13| 0g 9.99 14749 -6.49 131- H4 9:56 142-89 -1"89
a0 32 35. a -3-a - a9-6
G8. 15
31-30 -5.30 36 -3 -h3
11"a5 9G-a5 -10-a5
30. 18 0-a4 15. 15 79:0S 6-95
-418 55.59
34 231
- 30
-34-59 3a.3) -1-37 a8 89
a34-a3 -7-89
33 18) -3- a34
158. 1a ag.88 259 a49-29 -18"a9
167-71
11 a1. 6a 19:29 15]-49 a9-0ol 167-a1
O.38 H45) 19-79
3 a65 - aa.5I
349 23-91-1-97 15a1
a34 - a3 30- 77 6-19
6 a49-29 15-7)
 SsG N-v3)1 Ss6= 5636 Ss6 =2au5SSE =I153
SS T= &4al9 SST =8a93)| SST =82965
SsR =g Sa SSR =8069] SsR -S1811
R =0- 433) R'=0.99aq R=0.9861
$SR =SST-SsG

R
SS T

Since R u cloe to | in
N¡ oe acept thu as
Ane best
model amomg the
Bg-a ak fhimg derby
Predict the weight of the fuh al a ten of lengh
following data
Li 13 13 16 17 19
wi
12 1a-a5 12. 13 1S-a5 13-5
13-3) 13-25 13-75 13.
Here pyedict +he model au =a+bL
 Ihe mormd Eqne re

as
loa t 52364b = 131-37 a= 12-a3
523G4a +3-aa xob = 6961D46 b =1-13x161
=000o17
= la. a3 + o 00o I7 L,

13 = 1S.636

3 1a-a5 -010O 7
Ss = -w) = 15)849
122 0 an35
I6 13 SsR = SST- SS 6 = 6116y
O.7a65
15. a5 RT SSG
= 000)3
sST
13.5
1% 1337 Sionce R<l we voayt
13.a5 evise +he nodel.
13-75
a1 13
O695o
 Calculote SSE, sSR, Ss T amd g for the follousing
data.
a5 30 85
3a

Et a model of the form y

which the bettey ndel.

Follo oing tuo modey ere fit for the data.

1) = -35-2 + tolax 0) = O-0q4l

Idemi4y the bettey model amoTnq them.
S l n ; _ )

atb=y a-352
na tb Ex =Zy Ey=3og b = Hog
aEx +b =48o
= 315o 1=-35-a +4-1a

=-35.2+4-1a - (y-) (9-9)

32
15 a [Link] qtag-36
86
30
331-a4 %36
3-6
12-46
35

646-ah aHIss-s4.

ss E= (y-g) = 696- a4

ss T = E(3-9,) =nn6-34
SSR = SSE-sSR = -aogo6
SSE
SST
Ft a model of h e type y= A by 4he method
of leat Square for the dleta Hemce find the
Value of p
19
1 6 l 3a

19a05a 490458&I 3a44307 -4.443o7

6 24137s69 a8-aB849 -9-a3849
31
14091l4
3a
G4000000 37.34 -5. 84
as
134350 944X169 n3. qo6a5
Hn045 8%) 3a-44367 - l 4431
6a149go 1.31x1a9

3
=A
3

3
-a4 7951
74705 a9a0 a95
4= = 34l.1à38 = 56.0
n -35. nqnl

SsE = E(9-3) z 383-45an

SST

O.468Qog
 CxO Tn the modeling of bak sishing derby, the
data pertoining to the length
4he fuh eYe
cocigt of
fo lle wd.
I731 4139 4331739
8al 8G a31187aaa65
Follooing tuoo nodell coere it for the data.

Tdentity the bter noolel

Con putim q the co-elliei ent ofanong +hen toy
w= -74a5+ O00 43a? deleiimation Ce
SAm

-4.37 233 - 74786

3a.3¬a38 -7 9G666
31
13|4346a q.56 o38
3a -5 Bo a5 -70-86

a5 703n 338 15.¬5n5

11:37338
3a. 3n388
39 -a8000 58 1a8. aai4
33 a9.0066
157 qq094
a3 q6666 -\:4GG66

59.0o058 5449ya 16aaaly

lo a-3 6S lo.778 G.
SsG = -co) = aaus. a1%
Ss T = 5 ( )= 8a437.0%
Ss R = SST- SSG = 6.116H
R |- SSG
Ss T

Ss & = a38). 495

SsT=8ag53. 56
SsR =o566 07 R=0.q1 a

beat ngdel.