You are on page 1of 4

Gradient Descent:

Gradient descent is an algorithm that minimizes


functional difference between correct value and
predicted value of a function or rule.
In simple word:
Gradient descent is an optimization algorithm used
to find the values of parameters (coefficients) of a
function (f) that minimizes a cost function (cost).
Gradient descent algorithm start working on initial
value of given coefficients and start manipulation
of that value using top-down of bottom-up
approach to reduce total cost of function and try to
move coefficient value such way that they can
match to the accurate value of that coefficients
where prediction accuracy is near to 100%
Applicable to:
Gradient descent is best used when the
parameters cannot be calculated analytically (e.g.
using linear algebra) and must be searched for by
an optimization algorithm.

Titanic Discussion:
Section A:
In Titanic dataset we can predict survivorship by
considering below attributes with given
calculations:
1)Survived
2) Sex
3)Pclass
4)Embarked
Survived+Sex:UsingSurvivedandSexattribute,Wecan
establishedthatinthethiscreaticalcasefemalesgot
aprioritytosavethemselvesasbelowcalculation
stated:
TotalNumberofPassenger+crew(Male+Female)=891
Sex
Male
Female

Total
Number
577
314

Survive
d
109
233

Death
468
81

Survivorship
%
~19
~74

Survived+Pclass:
Usingdataofboththeseattribute,Wecanstatethat
passengerofhigherclasslike(1st,2nd)
Haveadvantageofbelowpoints:
1)Theyareneartosuchlocationfromwherethecan
easilyreachtolifesavingfacilities.
2)Theycanhavemoreoptiontosavethemselves.Like
lifesavingjacketsandall.
3)TheymayseatedinFirsthalfofTitanicasShip

Startedsinkingfrombackhalf.
Pclass
1
2
3

Total
Number
216
184
491

Survive
d
136
87
119

Death

Survivorship
%
~63
~47
~24

80
97
372

3)Sex+Embarked+Survived:
Combinationofthese3parametergivesusnumberof
femalesurvivedfromeachEmbarkedstationandalso
support1sthypothesisofthispaper:
Embarked

Total
Number

Total
Survived

Total
Female
Survived

Total
Female
Survived

S
C
Q

644
168
30

217
93
36

203
73
372

140
64
27

These are the attribute which dont help to count survivorship:

Name,SibSp,Parch

SectionB:
Ticket:
UsingTicketattribute,wecantracetheirrespective
probableclassesandtheirsurvivorship.
SectionC:
Topredictcabinmissingvalueswecanusebelow
parameters:

ByusingFarerangeandrelativecabinnumberand
Pclasswecanpredictthosevalue.

You might also like