DEPTH FIRST SEARCH

PYTHON CODE

s = (1111 2 2 2 2 2 2 2 2 2 2 151515 15 15);

t = [3 5 4 214 611121310 7 9 81516171918 20) ;

G = graph(s,t);

plot(G)

v = dfsearch(G,17)

OUTPUT

v=

17
15
2
1
3
4

5
6

7
8

10
11
12
13
_I
14

16

18

19

20

3.5

25

15

OS'----'----'----'----'----'---'----'---'
-2 0 2 4 6 8 10 12 14
ENSEMBLIMG TECHNIQUES

STACKING

PYTHON CODING

from numpy import mean

from numpy Import std

from sklearn.datasets import make_classification

from sklearn.model_selection impon cross_val_score

from sklearn.model_selection import RepeatedStratmedKFold

from sklearn.linear_model import LogisticRegression

from sklearn.nelghbors Import KNeighborsClassifler

from sklearn.tree import DecisionTreeClassifier

from sklearn.svm import SVC

from sklea, n.nalve_bayes import GaussianNB

from matplotllb import pyplot

def get_dataset(J:

X,y = make_classification(n_samples=l000. n_features=20, n_informative=15,

n_redundant=5, random_state=l)

return X. y

def get_models():

models = diet()

models['lr'J = LogisticRegression()

models['knn') = KNelghborsClassifier()

models['cart'] = DecisionTreeClassifier{l
models['svm'J = SVC()

models['bayes') = GausslanNB()

return models
_J
>knn 0.931 (0.025)

>cart 0 .825 (0.043)

>svm 0 .957 (0.020)

>bayes 0.833 (0.031)

1.00

0.95

0.90 ~ 0 ~
0

0 .85 0

0.80

0.75

0.70
0

Ir knn cart svm bayes

7
def evaluate_model(model, X, y):

cv = RepeatedStratifiedKFold(n_splits=l0, n_repeats=3, random_state=1)

scores = cross_val_score(model, X. y, scoring='accuracy', cv=cv, n_jobs=-1 ,

error_score= 'raise')

return scores

X, y = get_dataset()

models= get_models()

results, names = list(), list()

for name, model in models.items():

scores = evaluate_model(model, X. y)

results.append(scores)

names.append(name)

print('>%s %.3f (%.3f)' % (name, mean(scores), std(scores)))

pyplot.boxplot(results, labels=names, showmeans=True)

pyplot.show()

OUTPUT

>Ir 0.866 (0.029)

 /
1 ~oad f ishe ri ris
2 inds = ~st rcmp( species, 'setosa ' ) ;
3 X = meas(inds,3:4) ;
4 y = species( inds) ;
5 train_ feat= [meas ( 1: 40, : ) ; meas (51: 90, : ) ; meas (101: 140, : ) ] ;
6 train_label=[species (l:40, : ) ; species(51 :90 , : ) ; species(101 :140, : )]
7 test_ feat= [meas (41 : 50, : ) ; meas ( 91: 100 , : ) ; meas ( 141: 150, : ) ] ;
8 test_label=[ species( 41:50, : ) ; species(91:100, : ) ;species(141:150, : ) ]
9
10 train_ f eat= [meas ( 51 : 90, : ) ; meas (101: 140, : ) ] ;
11 train_label=[species (Sl:90, : ) ; species( 101:140, : ) ]
12 test_ feat= [meas ( 91 : 100, : ) ; meas (141 : 150, : ) ] ;
13 test_label=[ speci es(91:100 , : ) ; species( 141:150, :)]
14
15 ct reeModel = fit ct ree (train_ feat, train_ label) ;
16 predict ed_op=predict (ct reeModel,test _f eat ) ;
17

t est label=
20><1 cell array

20xl cell array {'versi col or ' }

{'versi color '}
{'versicolor'} {'versicolor'}
{'versicolor'} {'versi color'}
{'versicolor'} {'versi color '}
{'versicolor'} { 'versi color'}
{'versicolor'}
{'versicolor'} { 'versicolor'}
{ 'versicolor'} {'versi color '}
{ ' versicolor'} {'versicolor'}
{ 'versi color'} {'versi color'}
{ 'versi color'} { 'virgini ca ' }
{'virginica ' } {'virginica' }
{'virginica ' }
{'vi rginica ' } { 'vi rginica ' }
{'virginica' } { 'virginica' }
{ ' vi rgi nica ' } {'virginica' }
{ ' vi rginica ' } {'virginica' }
{'virginica ' } {'virginica'}
{'virginica' } {'virginica' }
{'virginica' } {'virginica ' }
{'virginica ' }
{'virginica' }
 ,
1 ~oad fis heriris
2 inds = ~strcmp(species , •setosa' ) ;
3 X = meas(inds,3 :4);
4 y = species(inds);
5 train_ f eat= [meas ( 1: 40, : ) ; meas ( 51: 90, : ) ; meas ( 101: 140, : ) ] ;
6 train_label=[ species(l :40, : ) ; species(Sl :90 , : ) ; speci es( 101 :140, : )]
7 test_feat=[meas(41:50, : ) ;meas(91:100 , :) ;meas( 141:150 , : ) ] ;
8 t est_label= [species(41 :50 , : ) ; species(91:100, :);species(141 :150 , : ) ]
9
10 train_feat=[meas(Sl :90 , : ) ;meas( 101 : 140, : )] ;
11 train_label=[ speci es(Sl :90 , :) ; species(l01:140, : ) ]
12 test_ feat= [meas ( 91: 100 , : ) ; meas ( 141: 150, : ) ] ;
13 t est_label= [species(91 :100, : ) ; species( 141:150 , : ) ]
14
15S\JT"l1odel = fit csvm(train_feat,train_label) ;
16 predicted_op=:predict (S\M-'lodel,test_feat};
17

test label= 20xl cell array

20xl cell array {'versicolor'}

{'versicolor'}
{ 'versicolor'} {'versicolor'}
{'versicolor'} {'versicolor'}
{'versicolor'} {'versicolor'}
{' versicolor'} {'versicolor'}
{'versicolor'} {'versicolor'}
{'versicolor'} { 'versicolor'}
{'versicolor'} {'versicolor'}
{'versicolor'} {'versicolor'}
{' versicolor'} {'virginica' }
{'versicolor'} {'virginica' }
{ 'virginica' } {'virgin ica' }
{'virginica' } {'virginica' }
{'virginica' } {'virginica' }
{'virginica' } { 'virginica' }
{'virginica' } {'virginica' }
{'virginica' } {'virginica ' }
{'virginica' } {'virginica' }
{'virginica' } {'virginica' }
{'virginica' }
{'virginica' }
>> svm
 I II

Bayes1an r egress1on:
1

f.rom1s;klearn1.dat.as,ets i1 m1p0rt llc a1d_boston

1

from sklearni.1m1odell_s,ele,ction ilmpoirt

t1 In_'test_spl li t
ra1
from1siklea,rn .mI etrics i1 1
m port r2_SC0're 1 1

from1s!klearn,. linear_model lm1port Ba'y'e.sianRi dge

1 1

da1tas et = lc ad_b1ost onl)

1 1 1

X, y = 1da1.as,et1data., 1dataset~targ1e t
X_train, X_Jest, y_train, ·y_test = train_t:est_spUt(X,,
1

v, test.......size = 0,15, ra1n,dom_state = 42)

(BAYESIAN REGRESSION:)rmo,del = Ba·yesianRidge()
1
modetflt(X_traln,, y_trainl
1
pr·edlict'ion = m,odeil'"predlict1(X_test)1
1 1

1
1
pri1n t(r Tes,t Set r2 score : tr2,_ sc01 re(y_test,
predliction) )1'11 )
1

Out1
pu't :
 Bayesian networks:

import numpy as np
import pandas as pd
importcsv
from pgmpy.est imators import
MaximumlikelihoodEst imator
from pgmpy.models import BayesianModel
from pgmpy.inference import
VarlableEHmination
heartDisease = pd.read_csv('heart.csv')
heartDisease =
heartDisease.replace('?',np.nan)
print('Samp le Instances from t he dataset
are given below')
prlnt(heartDisease.head())
prlnt ('\n Attributes and datatypes')
print (heartDisease.dtypes)
model=Bayesla nModel( [('age','hea rtd
isease'),( 'gender', 'heartdisease'),{' exang','h
eartdlsease' ), ('cp','heartdlsease'),('heartdis
ease',' restecg'),('hea rtdlsease', 'chol') I)

tun:U'.q CFD U1nq llu1&::1 llttl1t.ocd utlUtc:t

lnft:cc1nq v1ui 1a,u1.:: St:v::,1

I 11unl1lnu1 ~IUUtdl.a.,..t) I

I be1rUl.1cuec01 0.:0:2 I
I tt1n:t1111itl11 0.0000 I

I bl1:t411tu1 UI o.io:s 1
-------------I
I l:clrt:h~~t141 I O.HI!

----♦------·-
I tun4UtUt ;:l:1 (hu-..4uu..u1 I

I hlt:WtlUtfOI O.lC!O I
·-·--·--------
I 0.1:st I
htart:UltlltC!I

I hta:t:11J~l21

I tt1:tl112cue1J1 o.u,· I
I 1a1rw2eue111 0.:12! I
 6:05 PM GI tl/J $ ·· @ "' .,,11 C!D

MATLAB Drive ' multiple.m

l xl = [1; 2; 3; 4];
2 X 2 = [ 4; 5; 8; 2] ;
3 y = [1; 6; 8; 12] ;
4 n=length (xl);
5 xl sqr = xl. * xl;
6 x2sqr = x2. * x2;
7 xly = xl. "'y;
8 x2y = x2. * y;
9 xlx2 = xl. * x2;
10 sum xl = sum ( xl);
11 sum- x2 = sum (x2);
12 sum=y = sum(y);
13 sum_xlsqr = sum(xlsqr);
14 sum_x2sqr = sum(x2sqr);
15 sum_xly = sum(xly);
16 sum_x2y = sum (x2y);
17 sum xlx2 = sum ( xlx2 ) ;
18 y_bar = sum_y/n;
19 xl bar= sum xl / n;
20 x2-ba r = sum - x2/ n;
21 xlsqr_bar = sum_xlsqr/ n;
22 x2sqr_bar = sum_x2sqr/ n;
23 xly_bar = sum_xly/ n;
24 x2y_bar = sum_x2y/ n;
25 xlx2 bar= sum xlx2/ n;
26 sum_xlwholesqr-= sum_xl."'2;
27 sumr_xlsqr = sum_xlsqr-(sum_xlwholesqr/ n);
28 sum_x2wholesqr = sum_x2."'2;
29 sumr_x2sqr = sum_x2sqr-(sum_x2wholesqr/ n);
30 sumr_xly = sum_xly- ( (sum_xl • sum_y) /n);
31 sumr_x2y = sum_x2y- ( (sum_x2~sum_y )/n);
32 sumr xlx2 = sum xlx2- ( (sum xl•sum x2)/n ) ;
33 %%find b1\ 01, - - -

34 nrl = (sumr_x2sqr •sumr_xly ) - (sumr_xlx2•sumr_x2y);

35 dr =(sumr_xlsqr t< sumr_x2sqr ) - ( (sumr_xlx2 ) . "'2);
36 bl = nrl/dr;
37 %%find b2\%
38 n r2 = (sumr_xlsq r *sumr_x2y ) - (sumr_xlx2•sumr_xly);
39 b2 = nr2/dr;
40 °6%f ind bG"6%
41 b0 = y_bar-(bl• xl _bar) - (b2*x2_bar);
42 ypred = bG+ (bl• xl ) + ( b2"'x2) ;
43 scatter3(xl,x2,y, ' filled' );
44 hold on ;
45 plot3{xl,x2,y, ' r' . 'linewidth' ,2);
46 plot3(xl,x2,ypred, 'g' , 'linewidth ' ,2);
47
48
49
50
 38x4 table
Event Node Edge Edgelndex

start node 1 NaN NaN NaN

discovernode 1 NaN NaN NaN
edgetonew NaN 1 7 1
discovernode 7 NaN NaN NaN
finishnode 1 NaN NaN NaN
edgetonew NaN 7 3 10
discovernode 3 NaN NaN NaN
finishnode 7 NaN NaN NaN
edgetofinished NaN 3 1 3
edgetonew NaN 3 5 4
di scovernode 5 NaN NaN NaN
edgetonew NaN 3 6 5
discovernode 6 NaN NaN NaN
edgetonew NaN 3 8 6

discovernode 2 NaN NaN NaN

finishnode 4 NaN NaN NaN
edgetofinished NaN 2 6 2
finishnode 2 NaN NaN NaN
start node 9 NaN NaN NaN
discovernode 9 NaN NaN NaN
edgetofinished NaN 9 1 12
edgetofi nished NaN 9 6 13
edgetofinished NaN 9 8 14
finishnode 9 NaN NaN NaN
start node 10 NaN NaN NaN
discovernode 10 NaN NaN NaN
edgetofinished NaN 10 2 15
finishnode 10 NaN NaN NaN
-. - -- --
MATLAB Drive ) bfs.m

1 s = [ 1 2 3 3 3 3 4 5 6 7 8 9 9 9 10);
2 t = [ 7 6 1 5 6 8 2 4 4 3 7 1 6 8 2];
3 G = digraph (s ,t);
4 plot (G )
S T= bfsearch(G, 1, 'all events' , 'Restart' , true )
6
 • 10

•6 .. 4

,I
,9
'
I/
·•- -
l
~

"'
- • 3 -- 1
\ /1 ~

\I
~7
 SUPPORT VECTOR MACHINE
l
PYTHON CODE

Import numpy a.s np

Import matploUib.pyplot as pit

from skleam Im.port svm, datasets

Iris = datasets.loadJrls()
X = Iris.data[:. :2]

y = Iris 1a,ge1
C=1.0

SVC = svm .SVCtkomcl ='linear', C = 1).lit(X. Yl

x_mln, x_max = X(:, OJ.mini) • 1, X[:. OJ .max() ..-1

y_mln, y_max = X[:, 1J.mln() • t, X(:, l J.max() +-1

h = (x_max I x_mln)/ 100

)(J(, yy = np.mesrhgrld(np.arange(x_mln, x. max. h). np.arange(Y_min, y _ma x. h)l

pll subplot(l . , . 1)

Z = svc.prcdlct(np.c_(xx.ravel(), yy ravel()))
Z = Z.reshape(1CX.shape)

pll.contourf(xx. yy, Z. cmap = pit cm .Paired. alpha = 0 8)

plt.scatlru(X[', OJ, X[·. l J, c = y, cmap = plLcm Pahedl

pll.xlabet('Sepal length' )

plL. ylabel('Sepal Width'}

pll.xllm(xx.mln(), xx.max())
plLUtle('SVC with linear kernel')

plLshow()

OUTPUT
-.
 SVC with linear kernel

5.0 11

4.5
I,

~-0

5 3.5
;i '1
;r: I

i 3.0
I'

2. 5

2.0

L5

LO
6
5ep,1l l1!nQth
 K - MEANS CLUSTER
l
PYTHON CODE
import matplotlib.pyplot as pit

X = (4, 5, 10, 4, 3, 11. 14 , 6, 10, 12]

y = [21, 19, 24, 17, 16. 25, 24, 22, 21, 21]
plt. scatter(x, yJ

pit.show()
from skleam.cluster import KMeans

data = list(zip(x, y))

Inertias = O
for I in range(1,11):

l<means = KMeans(n_clusters=i)
kmeans.frt(data)

iner1Ias.append(kmeans.inenia.J

plt. plot(range(1 ,11), inertias, marker='o'J

plt.title('Elbow method')

plt.xlabel('Nurnber of clusters')
pit. ylabel('lnertla')

pit.show()

kmeans = KMeans(n_clusters=2)

kmeans.tit(data)

plt.searter(x. y. e=kmeailS.labels.J
pit. show()

OUTPUT

l r
 L

24
' •

22
•
• .. •
20

•
18

•
15 • .
4 6 I 10 12 14
J l
•
24 · • •
22 •
• • •
,o
•
18

•
16 • . .
• 6 e 10 u t•

Elbow method

Numl>ff or cluSle,s

7
ENSEMBUNG TECHNIQUES
BAGGING
from numpy import mean

from numpy impon std

from skleam.da1ase1s Import make_ctassificalfon

lrom sklearn model_selectlon lml)Ort c,oss_val_score

from ~klaam model_soloctlon lm?Ort RopeoIodS1roI1t1odKFold

lrom sl<loam.ensembte Import BagglngClassifler

X, y = make_classfficallon(n_samples=lOOO, n_features=20. r1.Jnfonnative=15,

n_redUndant=5. ranclom_st ale=5)

model = BagQITT(JClassffterO
cv " RepeatedSlrallnedKFold(n_spl tsa10, .,_,epeats=3, I-andom_state.,_ 11

11_scores = cross_val_score(model, X, y, scorlng-'accuracy·. cv-cv. n_jobs"·1,

error_scorc=-rnls-c',

prlnt('Accuracy: %.31 (%.3 fl' % (mean(n_scores), sl d(n_scores)))

OUTPUT
Accuracy 0.864 (0.038)