DATA MHNG -
AssiGNMeNT
Asuers
A) B(,8) c(s9) D:3) () t(as,30)
or poiut ,elustus ,rachak oits ano noe poiile
Data bts
A
3
C
9
25 30
Dstauee al poit vsong Euetideaw astaute
A B
A 1-4/4 2:823 485 9) 36400
B
5-67) -o1) 33-S)
414 25-SS3
2442
36-Yoo 34935 3357 253
Af-S (25J+(Som2)eB6t00
BCeJ3-2)+(-3)- 414
80J7-)-4(3-2)2 2 o)
AE 2J(g-)-+ (1-2)2 9-899
(2,6) A
H(4a) G(12)
fl6y (1.) ¬D($3) c(a.4) B(2,s) A(2,10)
Assar2
ED,
:A-B,
C, ptCon
Ppoj: ut Noise
beblu D,¬ e,A,poi:utReackabte
BA,c B CD
A’ B
Naigutoua:
<ES
Eps than les values Repaot
al
[Link](s-3)-4(30
1l -
24-142
25-Ss3I2s-7)-+(3o) DÆ5687 e
565 +(8-4)2
4J(-)-+(9-3)
19 DE J(s+(3o-3)
Mauhatan Dotaue
hroy A Clustu
A
6
B 6
4 3
7
6
lo lo
6
9
lo
H
3
2
Rkealulatig tta
Clustu 1(A)
A2, lo)
New tuctr 2,b)
foict : c(e,) D(5.) E(73) 64) H(49)
Nw cutie : 4$4796+y,
S
48454y89 z (&4)
Clutu 3 (G)
Poiuts : 8(2,5)
G(h2)
Naw Ccte : 24 ,sn (-5,3.5)2
(2, lo) (6i6) (s,3)
b23
0-23
o-38 o-17 o-30
o-S
p2-us
p3 o o17
o 5 l7
2
9
Nestco Csta"
Adabasct fousa sw nisclasitud pict . Msclasjed port
more ou thue
diyutt poiuta in next taactou .
lower wigkt
Casitatio talyts
Poiut 2 (o) : locatzal on Highe
folat 3 lo): locatd on ngut nls classifrcd
Point 4 (0) : beatd u
Petut s (A): locadkdl
locadad onon gut dlasitd tornoty J lower werglt
metuol tedt builde
Random forut s au thtube karninthor otput to ugroue
muttl dedsbw trs adl tou bins bekind Raudb foret
clastatib aturay. The ky ida de'solwtre iarodauig
is to reoue the ntue-oinoliohual
rancbmres n botuthe traidng data Quol the fastse seestiow
rocs
Suys:
" Gradt multipi boottiaprd sample fon ealuing datast by
Shstens wtw repleemut.
2- Build Deusow trce
. for eagh
beotsbappa sauagle ,traiw a deuiabu tre
" AA eacw ttctitbtu irtunal noole , seleet a randonw
Subst
tcatura to toMsdw for the but spb,
3- redtctiow
"for clasifiestbw, eauhw te output a clag tabel ; the
all trees .
for vgrasou, the prlidtbna Cue
4- Repeate and t9gat
tu to
Jmpat ttikte Sleetiow at Jrtnal Noda i
Trg : Dikuat tnu ltars fow Vamea
Crates Dtde modlel robeytnes
featue Subsea
irreltvat cus , lading tb bettu petormanee
’ lagmovu
Qway:
’ Reduts lorrlatiow amang tres les torrelatiow leask to betu atraging
paditons Overall model
i) Vavavitric
’Assumeo that tte dato ousbutiow
’ Dolimd by a hcd umba d paramgtus
’ eg: GM, ligite ngraaitw
wheú the datu outhbtiow well- knosw
Non- Paramctic
’ Model tan grow with the sze
’ Moe tuab Vand cae kandle uglw dkt thne data
’ g: k-AN, DESCAN ,lselactiouw forat stucara
’ Sutab or seMaÝOS whue data
olishibustow s vuknown or
Srmgular
’ Tiains mukijk maduls ndguduty ou
boctshad aast
’ Redluis varñaute and prat aasftig
’
’Sequoataly trains ak moded , cach ene
trtos madde
Cases
’ Redua aa vaiante but can
polat
poiut wite ateadt wipts wighbouu wituin ps to woeet
->4 forms the catal pat aclust
Notse poiut
’+ poiut that doa nat hae eouglw neighbours wtuu
to meet the niiusm
poid Vguinaut
->t doa hot belong to any tluytu and is considued
butier or hoise
Bordu þolut
polat -het ha fer toan niapts ndghkous wh
¬ps , but is reachable om a Core poit
a clhst
i lassiftow Onated Measue
lugas chutaing rauk to prdefincd labeli or broaw
eva luata how well the clustu altgn
witw knon classes
claHu
ey:Preistou, reall , 2 sore, and Jnalex
’Usd whehv true clas labels are Queilabtefor
evaatow
Sinilaity Oristd Masuru
-’focusy ou intnal clustu Suh as cohesivenes and
Sepuatiouw
>Na veed Hor prede7ined labels
8 thote &ort , Dumn Jnlex , Daviy - loudin Snox
Sita b whew clas labcls are Uuknowk and only Sin lai y
betweew polats mattus.
Anser
datast with vaging domsiea ,,absote distante or
dosty beud ommamaly dehutons may ql. tor g rgfonu
witdense and spare
Sin a
clustta ,a
pa-se rgion might not be anamolous , olâk a shlar
ancnaty
Suth casu, relatve Qnomale swre re Vesuctia! as
polts locad bohavior to that
ahomales more
I Relate astane based soro
Stont
uith
Comparu the asg clistane fom a poiut to t* nejghbors
olistane Btweeu temselve
Store >) reltatu an
2- Relatie Density basco Scort (L0r)
ananaly
The lacal oadtfactor(10f) tongara densty apolit utu
neghhou
local rachabuhy diny P
Avswer 6
clutu i
mi
class
Mi: No olcbs
-4
676 696
B93 613 63
ejz -ol36 + ool.36+ o o5+ o' o43+or035
2
Ctotal z -354 leg,2 -34| lo4. 34| 923
3204 V 3204 3204 32o4 3204 320 4 32lo4
-973, lay2273-738 lbg, 332
3204 3264 3204 8204
L, 4.,674
Mar
Mi
093
pl2 6-529
Ouerall
613 * o975 1562 os29 + 69 9 eo49
3204 3204 32o4
Custu
O202
2
3 o-529
049
Tote |-44
Cust labd
P2. o7
2
Matn
torrlatiow betoew
thd uetor nz <o,0-65,0S,o7,06,0 3
and veter y2 <l), , o O, o,>
motni onl ideal 'matix
6
&s-b- 6)+ (o- -o6)+ (o7- 66)2
6 (o3- o-6)+(o
+lo-6-o6)+ (os- 06)
| 6 ol703
2 toto+o+o+ e o33
6
+Q-od3 )2
lo ly) (o-3-6°6) o)+(oss-o)^le-o3)+
(o-s-b6)* (o-o33) + (o-7-o6) (o-o.3) +
(o-6 -ob) (o-o33) +(o-sro6) (Iro3)
Cor (y) -o)
Awerlo Color leas teigut
4 Short
2 Gren Tal! M
2 No
Grew Short Yy
Shot M
2. Short No
2 Tall
whit No H
2.
Tal No
2
Shot H
yu
H
Tat hstaue
Boctalsap saula
O43,9 , 2,6,6,3
s9, 8, 64,7, 6,
s24,?,1, 8,SR,6
tugutT , lgt 2, smelyN , Colayz 4 M
Rule 2 Smellye N, colore W
higut T legs 2,
TytT, legs 2, smely eNo, tolere J-H
Mode! -H
Modele’H
Model3H
Sauple 1 nele
223j Sauple ’No matcldg rule
a
SauglesNo 00B seple
for wRandom forat, we randonay Seleet 2 out
4 aweilab featurs
I ’ $4, !,3,7, 2 ,6, 6,2
2’5, 3,3 , !,4,7 , 6,5
3 2 , 4,7,!,3,S6
Roncom forst Deuttn Rula
Tre 4 (Bs) : 4,'3,7, 2,6,6
Re’ legs z 2
Tre 2 (ES2) ’ s, 4,7,,4,7, b,5j
Detault M
Tree3 (Rss) ’$4,417, !,3, S,9,6
olore wlutt Smellye
R2-
-’ lolore kolite Smely
Clesrhg st latanee
Modlel|- H
Madd 2
Model3 - M
00B Eor
Saugl Ao rule 0
Sufle 2
Sauple None
Mocl 9oB
Kandome ort oeg <