Download PPT

Telecommunication Case Modelling –
Call Center
C.Chudzian, J.Granat, W.Traczyk
Decision Support Systems Laboratory
National Institute of Telecommunications
Warsaw, Poland
The goal
Selecting prospective clients for targeting a
marketing campaign based on the existing data
(call center data, billing data and others).
Data Mining in Practice, Dortmund,
18th February 2003
2
Model of the call center
List of
the clients
Call center
Outgoing
script
Clients
Data base
Incoming
script
Reports
Data Mining in Practice, Dortmund,
18th February 2003
3
The data is distributed
Target group of clients
Analysis
Switch
Call center
Data Mining in Practice, Dortmund,
18th February 2003
4
Process of building a table
for data mining
Call
Center
Billing
New
attributes
Data
aggregation
SS7
New
attributes
Data
aggregation
Service
data
New
attributes
Data
aggregation
Synchronization
in time
Data mining
table
Client
data
Data Mining in Practice, Dortmund,
18th February 2003
5
Classification of clients
Switch data
Billing
YES
No
List of clients
Call Center
Data Mining in Practice, Dortmund,
18th February 2003
6
SAS Enterprise Miner
Data Mining in Practice, Dortmund,
18th February 2003
7
MiningMart requirements
Operator
Concept
Concept
..
1
Column set
Data
source
Operator
Concept
Table
View
Relational
data model
Column set
Table
View
Relational
data model
Data Mining in Practice, Dortmund,
18th February 2003
Column set
Table
data prepared
for mining
View
Relational
data model
Data
prepared
for
mining
8
The set of operators
CDR
Client
Call
class
Time
.........
Client
Data Mining in Practice, Dortmund,
18th February 2003
A1
A2
.........
9
Preprocessing (part I)
Process details
for all clients
SpecifiedStatistics
Client Features I
UnSegment
Features I for all clients
Process details for
a client
Segmentation
RowSelectionByQuery
SpecifiedStatistics
TimeIntervalManualDiscretization
RowSelectionByQuery
SpecifiedStatistics
Client Features II
UnSegment
Features II for all clients
JoinByKey
Data Mining in Practice, Dortmund,
18th February 2003
Client Features III
UnSegment
Features III for all clients
Features (I,II,III) for all
clients
10
Preprocessing (part II)
Features(I,II,III) for all
clients
GenericFeatureConstruction
All features for all
clients
Call Center
Data
UnionByKey
UnionByKey
Service users
Service info
GenericFeatureConstruction
Mining Concept
Service info (switch
prefered)
Data Mining in Practice, Dortmund,
18th February 2003
11
NIT case - HCI view
Data Mining in Practice, Dortmund,
18th February 2003
12
Business data
Data Mining in Practice, Dortmund,
18th February 2003
13
NIT case – preprocessing steps
Data Mining in Practice, Dortmund,
18th February 2003
14
Conclusions
Conceptual modeling improves:
• the understanding of the data preparation
process
• maintenance of the data preparation
process
• Knowledge transfer for other people
We do not need to use programming language
Data Mining in Practice, Dortmund,
18th February 2003
15
Questions & discussion
?
[email protected]
Data Mining in Practice, Dortmund,
18th February 2003
16