BUSINESS

FOR
CASTING
Btl
Principal )
(f)
 Type of algorithms that can begrouped in 3

categories :

D supervised learning -

Supervised learning
algorithms are trained using dabble data .

e.) Supervised elearning model takes did eat feed

back to cheek
if it is
predicting correct
output 08 not -

D Supervised
elearning model predicts the output .

9 In
supervised learning input data
model
is
provided to the
along with
output .

9
Yhegoalofsopuised learning input data
Bromide d to
is
model with the
§ Ten
along output
goal of supervised
,

the
elearning is to Isa in
model so that it
it is given
can
predict the output when
new dates
,

⑨
supervised learning can be
categorised in
classification and
Regression problems .

a) Supervised
learning can be used
for those cases ,
where we know the
input as were as
eosrespoing
Output . and supervised learning provides accurate results .
988 includes various algorithms such as linear

Regression logistic Res session suppose tveotos

, ,

Machine ,
Multi -
Chess
classification , Decision
tree
, Bayesian logic etc -

D UNSUPERVISED LEARNING

•) unsupervised learning algorithms are too in ed

unitabled data
using .

d) unsupervised learning model does is 2- bake

any
feedback
a) Unsupervised learning model finds the hidden

partner n in dates .

D In supervised input date

deworming •
my is

provided to lbhe model .

d) Thegoal find
of unsupervised learning
is to
the
hidden
pattern and
useful insights from
unknown dataset .

5 Unsupervised
learning can be classified in
clustering and Associations problems
.

It includes various
algorithms such as

clustering Ann and

Apisoii algorithm
D Gh is
.

may give less accurate oeesults as compared

tosupervised learning .

9 REINFORCEMENT LEARNING
works based on an action - brew and principle .

An
agent learns to reach goal by
iteratively
the new and of its actions .

) linear Regression
•

linear ee
egression is a
supervised I earning
algorithm and Isi es to model the chelationship
betwn a con ti nous target variable and one or

more
independent variable by fitting airline as
equation .

In simple words it ties to find out chelationship

beacon DU and IU

as IU grows up Dv grows as well .

formula =
a-
-

Gy )§✗Y -

(G) (sty
rice xD -

(exp
b- - •
(Exy) -

(Ex) (Eg)
)
-
n
( e E)> -

( ex
Support vector machine is used to
find out
on lab bleed data .

Thegoal the SUM to createbee best

of
is

08
line
decision boundary that can segregate
osr

n
dimensional space into class us So that
-

we can
easily put clue new data bointin
Are
correct
category in the
future .
5-
¥ >a

{
fraud detection .

diseases diagnosis
Emergency
"
detection
Spam " No "
spam .

These are the

places
-

where
you can see
logistics be
egression .
MEEDOFLOGISTI-REGRESSF.co#
When even bruin is categorical data ie -

Yes , No , True , False 1 ,

0 then we have to

Use dog is tie regression -

Classification of logistic Regression

y =

Sigmoid function .

mania bke
z
=
Independent .

= 2.718
e. =
Euler constant

output
y =
.

Sigmoid function is
just trying to convert

the Independent variable into a expression

of
probability that change bebwn
I with
oeespect to 021
Obe 1 hi hota hai
Logistic messes ion main .