2Deig Linean A

lttoatin
Fornstps? which haig 8reptestntvaniabl
as: d tplan
hight thcaleulates dlape onvaniable depr
e b hetght intenepttndepndent
b each to
y, with the
O,
dentCpndictors). co
b Input, whneels0;
e Nynesstn4" (2). betn a
t t thu
porcaptthsileanyengeon 2-5(6)+30 m we vaniabte
) olp ket' 5 (the ywont
taleulate p initi anaytny
1
either mate
ort to dependent
It
else wughb prtedtet asums suitabe
3ith
: 42.5
0 %paat
o statstieal
t a
yugt vámiable
outputs 1. with datc
2-5%+ s
k chany person' mdatonihiplinean
two on (tanget) Marmple
me
0.value 30 thod
AND terahons. ,
hight
thun with
wdght and usc
haght)weight,and tegnsson
Linean
" one modu to
Tnuth Considen g vaniables
bt?
a and besed
gneatnthan person independent
mone
tat ()table the
cl ne
the
bast an on
pencepkon line thin
twigh).yncton
ith
 o, ) b) a)
w, b w,
E-0+ 3lhput w,=05 lngut b
= =0-2 =05 W ylnput
05 o.2t
;(05() (,, o 0-5t (,, 04 0-5 o5
t
o.2(0-
+0-)o.2 +
o-2( O"2(0- o.2(0-1) t (a,,)
)(o.1) (o o-2(o-1)0 lo)
+ ) 0-1) +
) (0)ot 1) i)-
(o) (\,o) o2 olo)
(1) (o) (o,
() +
o) 02 o.y) o)

6) a) \tratton2: b w, ) b
0-2(0-i)0t
b= =-0:2
8o5 nput =05
Input 0 i Input 0+
D5
-02
(,,) 0.2
0-5 (x, + (a,,
(o) (o) + (0-)
02(|- 02(-
+ )= ,)
o +
(1)+ (o, (o) (o, bzo 0,
1) ot o) 0) o) ()
(o 1
o)
2)D

d) trtt+) b-0-2= ,
b
lnput
9 tnput&} deetoto-eH
l8nput -02
-o-2 *}lo-3) 0+ 05
tot)+ott)+t + o.2 +
( t0:2( o.5l1) (, 02(o-
o-2lo
(0-
aa)- + 3)
b-1) -0) 0)(-0.2)
o)(o.2)
(l,1) o(o) (,o) =
+
(-o-2))

E A ihe
Py|)> I,5 mag two
untbased à A bnay madu to
a emanples Tknahibru
bepencaptnon
tostum bank vantable pnedtt st o
b
ame (pana a needd
went depen
ha mts stais Additnath
wtlt pmobabilibieo rdependnt
dnt
bleed pn do
to ane t ea tant
de7aultpnedt leannd
ndeendent
tanametn method
and +5) , g coneyaTtnats das
that aiabts
vaniabe
on
laun wh jom ws
loan a the
then l
dat (ojg. ard
a)
 wlth nihad seeda ) Object 3:
penfonn k -meanA
cduakny
Drstanee to (4, 6):
(12, 4) lle the steps
dusten
Y- co0ndnat 4
Distance to (12,4)
2

3 d) Ogeet 4:
Dstanc to (4,6):
S36+ 4 -(40
J(4-1o) +(G-4)?
12 cdus 2
Distance to (12,4):
Initial seeds ( centoida
(4,6) e) Objet 5
Cust 2: (i2, 4) Dstane to (4, 4):

each sbject the cdoNeat duatn (based on Eueiden Asgnd

cdustn 2
distance). Distance to (12, 4):
J02-12) 4(4-4)2
tucltdean Dstance fonmul:
Sttp 2 Updat the centnodo by calutety the men o the coondnats

Calcuate distanee pn both centnoido: Objeets q3ignd to dusten i:

a) objeet 1: 2+ 4 t6 =4 -'4t 6+ = 6
.3
istance to (4,6J:
dustu 2:
J(4-2)? (6-4 = 4 t4 =Ss
dushen 1
Distanee b (n, uj: 2 2

Check n convyn
b) et 2: a) Obect 1:
distane .ko, (4,4) : Drstanee to (46):
J(4-4) +(<-c)? =o (4-2) + (6- 4y2 Jaty
clushn'1
SS Assigned to
Distane to. (12, 4):
Dtanee to (14, 4): clustun 1
JS
Ji24) +(4- 6)
 b) Dbject 2: The entrioi da didn't horye a n the knat on
th 1St tenatton, the
Distanu b (4,6) Stnce, the centnoica are

And dgonithn haa convnged.

Dhtonee to (t1, 4: 5 The mojt cocnmoy inan nentaion
modl Mean Squane Ernon (MIE). It rneasuns the avs squened
dilerene beth th achual (obsenved) valus and the pnediced .valus.

) Object 3: The goal is o miotntse thin duniy tran

Drtanee to MSE (á
Distane (t, ):
where n n. data points (sampiu)
Tth deta point
i actual (tnue) velue om the
or th Th data potnt
the digenwe Btn the atuel
d) objee 4: The bnmule calculats the q:
Datane b (9,4): vae Cy:) ant the pinedikd valut ) en ach date pånt

Aitjned to
Drstance o (11, 4): cdustn 2 Psints A(3,1), B(4,6), c(s.s), D(<,4), ¬(%,3), F(«. 2),
Jln- io) (4-) JI +o | G(4,2), H(8,4)
Eps 2-5 (the nadis oh thu naghbonhaod ean en)
e) Objeet 5: Min Pts 3 (min. no. of poinb n the nigibounhond to be Constdd
Drston u to (4,6): potnt)
Step 1 : Caleulatk the' Btt Eudtan dittanee bet" h pin oh points to
Dstance to (i1, 4): duste 2 identM nuzhbouns othin the eostlon radBus (Eps):
(- 2)2+ (4-4) : JT:= I
custen 1:
a) A (3, ):
Disi. to
2 +4t 4 = 4
3 Drst to Negtboura Eps 2.5 ) 'B
Drst bo D: 4. 24 A cOne potnts (as MinPt 3)
Obj asigned to duster 2: Dit to E: 5. C5
4t4 =4 Dist to F!5. 83'
2

Dist H: 583
b) l4 ,6): )F(6.2)
Dtst. Ao A: 5. 83
Dist to A(3,) : ):41 Nughbounu ithin tp 25;
Dit 4.44
Dryt. to c(5,5): 41 D, E, 9
Dist. bo D(6,s): 2.82 oithin tps 2.5;, point

Dit b E3): 4i24

B aone pont ( as Min Pt 3)
Drst to (6, 2): 4.47
Dist to (2): 5 Dist. to 2. 82
Dit- to ($,4): 4.44
a) (7,2):
sithi
) * c(5.5)
D, E, F, H
Drst o Nlghboung within Eps 9.5 core point
Dst 2.23.

Dnt t ¬ :2-82 ot Drst. to

ore -potnt ( ao Minpb 3).

Dist st, t t: 2.23

Drst- to t: 3.i6 h) tlS,4):
d) D(6, 4)
Neshbouna ithin tps 2.5;
Dot: to A: 424
withtn i tps = 2.5 D, E, G
Dist to
Dist, b , : 2 cone point
Drst. to C: |. 41
D a cone poiot
DrskfoF 2.82
Dit to 42.23
Dist o q: 2.23
Dt. to tt: 2 (ORE POINIS : D, E, , 4,H
DUTLIERS: A, ,C
e) E(1, 3):
DisA o. A: 565
wIthin Eps 25)
gbuns
Dist: to C: 2. 82 D, F,G
Dst bo E a
point
Dist to F: 2
 1toation 1:
Input Dist Neanst
layn Output Point Dist. to Dist: to
C3(8,4)
clust

lagen Ci2, 10) C2l2,5)

Multt- lan A1 (2,18)
843 C

Pereeptnon
(ML) 6.08 C2
A2(2,5) 5
C3
A3 8,4) 608
OPERATloN: 424 5
A4(5,8) 3.60
a) FORhARO PROPAGA1ON: Input ani att multlied by whts , biass ,ene
added, and actvains each sthot 5
ane A5(5)
lay 4.12
C3
4:21

b) Loss CALCUTAT1ON) The model compane pnedicked outputs to átial' valus t.28 C2
A( l, 2) 6.06
wty a loss . (MSE)
As( 4,9) 2.23 4.47
) BACK PROPAQAT lOAN: The madul adjuot wyht ond blages ba pmopyaiy
bakand
Centroi da
Step 2 Updete
d) 1TERATION: This pnoess Hepeats aenos multiple epochs unki- h a) Custn 1:
minmid and Hhe model leanna Kkeetiveg
M, = 2+ 5 +4 = 3i61

An MpL a
Centnoid 1 : 36t, )
yncasion tasks, ett,
POrNTS: A(2, to) A2 (2,5), A3 (87 4) A4l5/8) , As(75), b) Custon 2:
2t | = 5 5 + 2 = 3.5
A6l6, 4), All,2), 2
2

Let elusten 1: A1 (2, 1o) Centoid 2 (. 5, 3. 5)

lusten 2: A2 | 2,5)
c) ustn.3:
custe 3: A3 (6,4) 4:33
a3 4 5+4
3

Skp 1: Asstgn points to ncanut (antnoid wst Guclidean distan

Centoid 3 = (45 4.33)
Stp3Reessign points to neaneot cetnoid ftoal custens
a) Cstor 1 AL, A4 A8
Points o C2 Neanes t b) Uustu 2: A2, A7
cluster
(3.67, ) (|.6)3.5) (, 4.33)
c) Austen 3
A3, AT A
Input
A1(2, to) |.94 651 q55 (nput
Sumnaton
A2(2,5) Output
5.04
A3\8,4) 6.61 G51 .o5 C3 Ackvatton
Auls,s) 54 4 14
As(4, 5) 5.20 5. 4
3
AG(6,4) 5.51 4i52 Bias
c3
All,2) i 58
Sirgle layen Poneptnon (sLP)
Asl4,4) O.33
555 C OPERATiON:

a) Custen 1! a) lnput: Paovtde le valus

, = 2 + 5+ 4 = 3. 67 6) waght maltplicaton :' Multtply ach e with nupeui ve' daght

2 c) Sumenation Add the weishtd lps and bias to get y u n Z +5
Centmold 1 = (3 6, 2)
b) Cuustn 2. d) Actvoaton: Appy thu

2
j22 5t2 35
2

Cent noi d 2 (.5, 3-5) ) Output: he pencephion produees binany casi jteton Mesult

J3 4+54 4
hek (a) 2 3
2.6 3.2 2-8
Centnold Y (4 4i33)
Sals (y) |2

Stp 4 : Check Fonmula uncan Regnession:

n conveyene
The eluchn acstgnments &ne
unchanged iom the pnerious tenation. straght ine i:
achivd. Convemgeace
 m
slape o lne Sales or week :
04G(4)4 0-94 416

Saes or wek i2:

Calculal-in th mean and

Clusten 1: ( , )
5
Clustur 2: (5,)
tuidean distane:
the distane bet dustos ond potnh usty
Calculating
et Calclatiy the slope (m: Cuustn
Dist to Dist: to
Potnts neaer
cil1, 1) C2( 5)A)
m
I(- 2) Cy-g) (reconds)
421
R (I,)
m= (-3)(1.2- 2.32) +(2-3)(18 - 2-3) +(3-3)(2-6 - 2.32) + (4-3) 6.10
R2 (15,2)
(3.2 -2-92)
+ (5-3) (2-8- 2-232) R3 (3, 4) 3-G 3.6

(t-3)2+ (2 -3) + (3-3)+ (4-)? + (5-3)* Ry (5,9) 72| C2

Rs (3:5,5) 2-5

R& (4:5,5) 5.31 2.06 C2

Ra (3+5,45) 4.30 2.91 C2

Calouletiy inkncup) (el: Points asslqned to dustr 1 :

(-25
+|·59

>C: 2.32- (0-u6 X3)

Points asstgne to duston 2:
4, # t5 t 5 + 4 5t4
=S35

Centoid 1: tt+3) l-25,1-5)

Recaleul at the ditance points and

Potnts Dist t
(Recoras)
Dist to. Nean wt
clust
Caleulety mean

40 +120+ 150+100+ 130

4 2 i18
C2(t5f) 5

t 055
Caleulaig rneen

R2(1-5,2) o55 3.92 y= l000 + (3004 1800+ (200+ 1380 = 1336

R3( 3, 4) +4 42 C2
Cateutakig the slope (ny:
R4( 5, ) 665 2-19
I(;-) y-g
Rs3 5 ,5) 041 C2 (%;-)
RC 455) 5 m (20-9 18)(1000- 1336) + (2o -e) (I300- 1334) +
R(3-5,4.5) C2
(150-116) ( 1803 -1336) +( (00-118) (n00-i336) t (130- 2)
(1s0-1336)
Points asigred to clust i: (40-118)+ (2o -te) + (15o-11a)+ (109- 1a) (133-i&)
, |+[5 = i-25
2 9408 t (-42) +14848 +2448 t528
4844 4t 1024t 324t 44
Centnoid 1 (l. 25, |-5)
2416 =il9|
fotnts assigned to duaten 2: 2280
, 3t5t35+ 4-5 3= 3.9 4 4+74 5+5t 45 5:
Caleul atiry the ntnupt (coi
Centrold 2:(34, 51)
As the centroid atten the a
tenatin enains unchayd, ’c= 69. 3g
hene, the agonthin convenys.
Now, or the advent'ement o $200, the salea shoutd be
220
Advertisetment (x) 20 50 13o

Sales cy) 1000 1300 1800 |200 1380 y l|91 (20o) + 6 38

* 245l. 38
15 Reoult
Prcdicted: spam Predd! not scam Totat Maths Carn(v)putn
(x)
Actue! 100
fail
Actual : not spam 20
Pass
Total |0 200
Pass

5 Fa|l
TNC True nyathve) 20
FP ( False positive) $0 5 Pass

FN (Falsc nyatve = 0

Acunauy Total
40+ 20 =
200
055
alg tudean calo ulae ditanee bun point a
Pre cision = 40 o52 6,S)
TP+FP Dist bo. (4,3) (4-6) +(3-s) 5-38

40 (G,4) (6-G)+ (+-8)

TP+FN Dt:

FI -Store 2. Pnedsion Reeal 2 052Xo9 Dot to (515) J5-6)+ (5-9)2 =316

Pre cisiton t Recol 52+ Þ.9 Drst bo (3, 8) JG- s)? (s -s)2 = 2
Specipirty TN 20 The neanest points ( ort eghbous) ane pts 2,3 and 5.
Sine, the majonity o% the neaest nushtbo una have the lasi "Pas , the
09
predictd ewlt

Givn:
FRR (False positive tate)z | - speukcity |- o2 o& Points: AI = (2,10) A2 (2, 5) A3= 8, 4) A4- (5 ,8)
A5 = (5) A6 (G4) A? , 2) A8 (4, 9)
Eps 2
MinP < 2

Cateulairy the distanes krom dikbtnet pints uty

Eucidean dstanee onmula.
 21 At
point. cone2) Pts Min (as 4.12 46
a not A4
2) >ptMin (as A5.5 e2) Epe ithin Naboun 5 43:
point 24 4.A2:
Dit
NughbounA4 444 A4(5,8): fon d)
2.23 Al: om Dist.
4,9): As( fon h) 128 A7
538 Abi point A5141
2) Ptr Mn (as 5A4i
pa'nt A5
A 2 A41 6o8 A2:
unnajhbo
s nT 728 A3: &48 :Alton Dist.
2;tp:- ithin Nughboun 3l6 92: tA3(8.4): fon c)
8o6 Al: tom Dist.
(1,2): Aa n a)
38 5A8:
38 5 f\? |2 4A6:
(41 451 poYnt ot A2 5A5:
point cone a A6 412 A4:
naghboury 4-24 A4:
45A3,
tp2; wthin Njhboun 6.08 A3:
Dstam
2
t;pst? thin Nughbun 4.12 A2: AI:5
2| Alkon Dist, n2(2.5): r b)
4): AG(6, fon
2.23 AS:
5A8:
i
.21 A6!
point Cone a not Al
3.6 A4: nughbouns no
point 2 tpt th
Nghbour
A5
AG
A3, 5 A2
48 A3:
Nuybon Dist. 5 A?trom Dist.
4,5): As( fon e) AI(2,
10): fo a)
Come nts
Nohe poht A, 2 , A, A At

Jio 3l6 then

(one poBnts AS, A5, At

Noise ph A, A2, A,

points and nose points Hem atns Same