Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
10 views

Data Science Using R

The document discusses how K-means and K-medoids clustering algorithms can be used for customer segmentation in industries like travel. These algorithms group customers into segments based on behaviors and demographics to enable targeted marketing campaigns. The document also discusses how these algorithms can be applied to problems in the airline industry like flight scheduling and anomaly detection of irregular booking patterns. It concludes that K-means and K-medoids are versatile techniques that have applications in data-driven decision making and personalization, customer profiling and revenue optimization in the travel industry.

Uploaded by

aman kesarwani
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
10 views

Data Science Using R

The document discusses how K-means and K-medoids clustering algorithms can be used for customer segmentation in industries like travel. These algorithms group customers into segments based on behaviors and demographics to enable targeted marketing campaigns. The document also discusses how these algorithms can be applied to problems in the airline industry like flight scheduling and anomaly detection of irregular booking patterns. It concludes that K-means and K-medoids are versatile techniques that have applications in data-driven decision making and personalization, customer profiling and revenue optimization in the travel industry.

Uploaded by

aman kesarwani
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

CASe STUDY

’ededive Analyti boneg' customertawelig


ovenasl
expeienu mean and

K-mehods algorithm tehigue1,


mebiods
Hinswer K-means and K-ethee
k- clusteing
are
K-means
agovi thm uudin predictive analts and daba
sienei have varioup applicatong
the travel indust
dlnt industies
metained în
incdeng
your þrcuioug message
X CuSTOMER SE QMENTATION K-mediod
K-meang and
In the traved industy Cuitomer baazd on
Can be used e to eg-ment
behavioury and demagraphico
and
their proen tea,
înto gvou ,travel agnus
usbmery
Juttaing usto
Campaigng, toilor
Can oeate targeted mar Kehirg user intoatong
tvavel packoqes iand improve

Understanding
-eu Can
wtomr
Liss in
sgmunt lust
through u
develoing compcihve
bed on Regmenhs and marhe
ATRLIn NheES stliniing
OPERATION
Contet airnes thea
the olgon
ight cheduing
Can be wed to olimize
gnmentz and route planning
related to ight dolay
cteni data
Cancellationt and
psenger ,prelerenA1oirlines
a nd

Can enhane ofentiond etiieny


travl enpeien
improue the ovenal
ANoMALY DETECTI0N identit
Both ansnalies
algorithmor outiers in ta Fo
in spotting
Can
the travl induty thisboo kig bahaviong
irreqlar patenn

(ONCUSION

kmeang K-mediod are vosatile cueteing


technigues that she ind apicationg in
Man Dabect o' the travd indurtay
onti ou to þenzonai zati'orn, Uustomen
butig ond rev en ue opimizohon
Sotiachon
a chieed

evaluation and vaidation


Boatstra method? vuampli
Ser The Boststrab Hethod is and
a vaikdaton
.Herc's
Rvaluaton and
tehrique uud tor
in breditve anal yhes and
data Beienu

hous h achived.
ReSAMPLIN (with
resamþlinges
Firt a .rom te or8
genenated
eplaconent are mesamþlea
he ome

ane
-nal Thest
a 2et.o>igina dodoat
ze the

your þedictive
for each vesanple 1Hou obly algorithm) to
a machine learning
modle' e
data. This mean
the resanpled
the model to ebampled data.
y STAS TICAL ANALYSTs.
modl to each resamþle
Ato iing the ohich mayincuda
mauy
you colect the rout
model ara meter, predi ctions or evaluaton
netia iKe ac ur a cypreusion or recall
LASTING IMPRESSIU

he
UALUA1ION AND VA LIDATI N

eti Ugted rehty rouide a betr


extiPng
mate themode's podomanu
ou undlenctand ity variabi
obustnes.
mode evation This Can be used tor

eimarg
and validaton inelud
enor rates
ates aelig
tabi nd
aMeaing model
bulding conidence irtenva'a
mode
rman
CoNLUSON
he Boot stra
hou method is vaduoble for
your þredicive model
fneozas t neo data and for
underct
4he
Jur model, nti
Un
nt Usouate with

parti cudarly usuanormane


when mei:tiz
you have limited
data uhen you uoant to avoid

You might also like