for Predictive Analytics in

collardsdebonairΔιαχείριση

6 Νοε 2013 (πριν από 3 χρόνια και 7 μήνες)

72 εμφανίσεις

ICMIS
-
2010

1

Knowledge Management Approach
for Predictive Analytics in
Marketing DSS using Temporal
Data Mining techniques


Sunita Soni, Jyothi Pillai,

Department of Computer Applications, Bhilai Institute of Technology,Durg

Dr. Ranjana Vyas,

MATS University,Raipur

Dr. O.P.Vyas


Indian Institute of Information Technology Allahabad.

International Conference on Information Security & Management of Technologies in Business

ICMIS
-
2010

2

Overview

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques


Introduction



Predictive Analysis in Marketing DSS



Temporal Associative Classifier




Knowledge Management Frame work



Experimental Results



Conclusions and Future Scope

ICMIS
-
2010

3


Introduction


Knowledge Discovery in Databases
.




Data Mining



Predictive Analytics

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

4

Knowledge Discovery in Databases
-
KDD



KDD

is non
-
trivial extraction of implicit, previously
unknown and potentially useful knowledge from
data.




KDD

is a significant concept related to data mining
and
business intelligence.


Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

5

Introduction


Knowledge Discovery in Databases



Data Mining



Predictive Analytics

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

6

Data Mining



Data mining
:

the core of KDD.


Data

Mining
:

discovery

of

hidden

knowledge,




unexpected

patterns

and

new

rules




from

large

databases
.


KDD

describes

the

whole

process

of

extraction

whereas

Data

mining

is

used

exclusively

for

the

discovery

stage

of

KDD

process
.


Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

7

Data Mining

in various forms is becoming a major component
of business operations.


Business processes involving data mining for
Knowledge
Management




Customer Relationship Management,



Business Intelligence,






Supply Chain Optimization,



Demand Forecasting,



Assortment Optimization,
etc
.



Data Mining in Business Processes

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

8

Introduction


Knowledge Discovery in Databases



Data Mining



Predictive Analytics

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

9

Data Mining Tasks


Common data mining tasks


Classification

[Predictive]


Clustering

[Descriptive]


Association Rule Discovery

[Descriptive]


Sequential Pattern Discovery

[Descriptive]


Regression
[Predictive]


Deviation Detection

[Predictive]


Prediction Tasks

-
Predict unknown or
future values of variables

Description Tasks

-
Find human
-
interpretable
patterns from data.

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

10

Overview

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques


Introduction



Predictive Analysis in Marketing DSS



Temporal Associative Classifier




Knowledge Management Frame work



Experimental Results



Conclusions and Future Scope

ICMIS
-
2010

11

Marketing DSS

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques


Marketing is the process of Identification of Demand and fulfilling
it in a profitable manner and in case of no demand or less
demand, it is the process of creation of demand and then
fulfilling it in a profitable manner.”

Types of Decision in Marketing (4 P’s).

1.
Product:

Modify the product in terms of size, quality, quantity etc.

2.
Price :
Cash,

EMI, Discount policy, interest rate in case of EMI.

3.
Promotion:

Advertising policy .

4.
Place :

Direct to the customer, Channel, Through retailer

ICMIS
-
2010

12


To take any decision regarding four piece of marketing ie
Product/Price/Promotion/place we have to analyze the environment
factors, competitors, consumer, past experience and performance.


Today , huge data repository are being maintained in every field
including business and valuable bits of information are embedded
in these data repository.


Because of huge size of the data source makes it impossible for a
human analyst to come up with interesting information(or pattern)
that will help in the Decision Making Process.





Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

Predictive Analysis in Marketing DSS

the Solution is Data Mining

ICMIS
-
2010

13

Data mining is related to building Decision Support Systems.



First, data mining can help identify relations and rules that can be
incorporated in Knowledge
-
driven DSS.



Second, case
-
based reasoning can be used to create a specific
Knowledge
-
driven DSS that can be used by a manager or a knowledge
worker who is trying to diagnosis problems in that "case" environment.



Third, data visualization tools can be incorporated with a structured
data set to assist managers in making a recurring decision where the
data set is routinely updated.


Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

Predictive Analysis in Marketing DSS

ICMIS
-
2010

14

Direct Marketing


Goal:

Reduce cost of mailing by
targeting

a set of consumers
likely to buy a new cell
-
phone product.


Approach:


Use the data for a similar product introduced before.


The
{buy, don’t buy}

decision forms the
class attribute
.


Use this information as input attributes to learn a classifier
model.


Collect various demographic, lifestyle and company related
information about all such customers.

Predictive Analysis: Application 1

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

15


Fraud Detection

-

Goal
: Predict fraudulent cases in credit card transactions.

-

Approach
:


Use credit card transactions and the information of account
-
holder
(
When does a customer buy, what does he buy, how often he pays
on time, etc

)as attributes.


Label past transactions as fraud or fair transactions. This forms the
class attribute.


Learn a model for the class of the transactions.


Use this model to detect fraud by observing credit card transactions
on an account.



Predictive Analysis

:
Application 2

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

16


By incorporating data mining techniques in marketing DSS,
retailers can improve their inventory logistics and thereby reduce
their cost in handling inventory



DSS with visualization tools may help to understand the composition
of the portfolio and help identify what changes need to be made in its
component stocks.




Predictive Analysis

:
Other

Applications

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

17

Overview

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques


Introduction



Predictive Analysis in Marketing DSS



Temporal Associative Classifier




Knowledge Management Frame work



Experimental Results



Conclusions and Future Scope

ICMIS
-
2010

18

Classification: Definition


Given a collection of records (
training set
)


Each record contains a set of
attributes
, one of the attributes is
the
class
.


Find a
model

for class attribute as a function of the values of other
attributes.


Goal:
previously unseen

records should be assigned a class as
accurately as possible.


A
test set

is used to determine the accuracy of the model.
Usually, the given data set is divided into training and test sets,
with training set used to build the model and test set used to
validate it.

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

19

General Approach for building classifier

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

Training Data

Learning


Algorithm

Learn Model

Model

(Classification Rules)

Test Data

Induction

Deduction

Apply
Model

ICMIS
-
2010

20

Association Rule Mining


Association rule

A


B,

where A, B


I , and A


B=

.


The

rule

X



Y

has

a

support

s

in

the

transaction

set

D

if

s
%

of


the

transactions

in

D

contain

X

Y

.


The

rule

X


Y

holds

in

the

transaction

set

D

with

confidence

c

if

c
%


of

transactions

in

D

that

contain

X

also

contain

Y

.






Strong rules are

the rules that satisfy minimum

support

and

confidence

threshold

values
and this framework is known as the
support
confidence framework

for association rule mining.

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

21

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

Association Rule Discovery


(Data Mining , Statistics)


䑡瑡t却牵捴畲敳


support=20%, confidence=85%

Supermarket shelf management.



Goal: To identify items that are bought together by sufficiently


many customers.



A classic rule

--


If a customer buys the book of ‘Data Mining’ and ‘Statistics ,
then he is very likely to buy the book of ‘Data Structure’ .


ICMIS
-
2010

22

Associative Classifiers

(AC)

Associative Classifiers

is a two step process :
-




CAR generation(Association Rule Mining)




Classification using CAR




Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

23

Associative Classification (AC) Problem

Given

a

labeled

training

data

set,

the

problem

is

to

derive

a

set

of

class

association

rules

(CARs)

from

the

training

data

set

which

satisfy

certain

user
-
constraints,

minimum

support

(
minsup
)

and

minimum

confidence

(
minconf
)
.


Common Associative Algorithms
:

CBA

: Classification Based on Association Rule Ming

CPAR
: Classification based on predictive association rule

CMAR
: Classification based on Multiple Association Rules

MCAR
: Multi
-
class classification based on association rule approach

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

24

Temporal Associative Classifiers

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

The

customer

interest

is

changing

with

respect

to

time

because

of

many

reasons

-




In

real

world,

the

items

have

the

dynamic

characteristic

in

terms

of

transaction,

which

have

seasonal

selling

rate

and

it

hold

time
-
based

associationship

with

another

item
.




For

example

during

the

festival

time

or

during

new

year

when

the

majority

of

customers

are

getting

an

improved

earning

or

during

the

first

ten

days

of

a

month

(salaried

employee)

have

the

high

purchasing

tendency



Also

during

the

depression

/

recession,

the

customer’s

investment

policy

or

purchasing

tendency

may

change
.





Also

some

items,

which

are

infrequent

in

whole

dataset

may

be

frequent

in

a

particular

time

period
.

ICMIS
-
2010

25


Adopting temporal dimension will give more realistic approach and will
yield much better and useful results in Predictive Analytics.

Hence, a novel approach of combining Associative Classifiers with
temporal dimension is being proposed.

The purpose of

temporal predictive system is to provide the pattern or
relationship among the items within time domain.

For example rather than the basic association rule of {bread}

{butter},
mining from the temporal data we can get a more insight rule that the
support of
{bread}

{butter} raises to 50% during 7 pm to 10 pm

everyday .

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

These type of rules are more informative and useful to

make a strategic
decision making in every field of business intelligence.


Temporal Associative Classifiers Cont……

ICMIS
-
2010

26

Overview

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques


Introduction



Predictive Analysis in Marketing DSS



Temporal Associative Classifier




Knowledge Management Frame work



Experimental Results



Conclusions and Future Scope

ICMIS
-
2010

27

T.A.C.

algorithm

Relational Data
Base

(Historical Data)

Prediction for business problem
such as forecasting the customer
Interest for given time granularity

Knowledge base for Predictive
Model system in business
Intelligence solutions

Temporal
Classification
Association rules

Knowledge Management
Module


Domain
Knowledge

(External
Resources)


Knowledge Management system using Temporal Associative Classifier

Methodology for Implementation


Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

28

Overview


Introduction



Predictive Analysis in Marketing DSS



Temporal Associative Classifier




Knowledge Management Frame work



Experimental Results



Conclusions and Future Scope

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

29

Experimental Result

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

Dataset


CPAR


TCPAR


CMAR


TCMAR


CBA


TCBA


Anneal


94.99


93.2


76.168


92.2


78.4


95.3


Breast


92.95


91.2


90.82


94.5


92.82


93.95


Heart


51.12


77.3


54.09


76.2


55.42


75.5


Hepatitis


74.34


72.1


78.33


74.4


42.5


77.5


Horse


81.57


88.3


67.47


89.4


56.16


79.8


Ionosphe


89.76


88.3


63.8


86.3


42.17


82.48


Iris


95.33


93.4


96


91.3


96


92.3


Pima


74.82


74.3


65.1


75.3


55.7


74.24


Wine


88.03


85.5


68.96


80.46


72.31


88.2


Zoo


95


93.2


40.36


92.3


60.18


88.2


Average
Accuracy


83.791


85.68


70.1


85.236


65.16


84.747


Comparison of average accuracy for various associative Classifiers

with their temporal counter part

ICMIS
-
2010

30

Overview


Introduction



Predictive Analysis in Marketing DSS



Temporal Associative Classifier




Knowledge Management Frame work



Experimental Results



Conclusions and Future Scope

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

31

Conclusion

1.
This

work

presents

a

framework

for

Knowledge

Management

system

and

discusses

the

result

of

Temporal

Associative

Classifier
.




2
.

TAC

is

found

to

be

an

effective

technique

to

extract

knowledge

from

the

database

.




3.

The model presented in the paper highlights the amalgamation
of TAC with Knowledge Management Approach.


Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

32

Future work




The model can be implemented to perform prediction or
classification in any sort of business problem in more efficient
and accurate manner which will ultimately help in decision
support system of any organisation.




2.

In future the model can be further explored for the given
application domain of super market sales to incorporate the
domain experience of related experts.

Knowledge Management Approach for Predictive Analytics in Marketing DSS using
Temporal Data Mining techniques

ICMIS
-
2010

33

Thank you