Using Text Mining to Enhance EnterpriseWide Business Decisions

1
Using Text Mining to Enhance Enterprise-Wide
Business Decisions
Online seminar presented by David Purcell
Doug Bryan June 19, 2002
2
Business Intelligence

• Get the Right Information
• To the Right Person
• At the Right Time
• To Make the Right

Business Decisions
3
Todays Agenda

• What is Text Mining?
• Structured Text vs. Unstructured Text
• Linguistics and how it is used in Text Mining
• Applications of Text Mining
• How Text Mining is used through an email
  application to streamline a business process
• How Text Mining can be used to discover patterns
  and relationships in text to streamline a
  manufacturing business process

4
Text Mining extracting business intelligence
from unstructured text using linguistics and
statistics enabling users to quickly identify key
concepts, and the relationships between them,
within thousands of documents
5
Text Mining Linguistics
The capability to understand human language is
provided to computers through the power of
linguistics - Natural Language Processing
NLP
Concepts
Text
.doc.pdf .pst .msg web
6
(No Transcript)
7
Linguistic Example
Insect
8
Linguistic Example
software
Insect
automotive
medicine,molecular biology
9
Today's Agenda

• What is Text Mining ?
• Structured Text vs. Unstructured Text
• Linguistics and how it is used in Text Mining
• Applications of Text Mining
• How Text Mining is used through an email
  application to streamline a business process
• How Text Mining can be used as an added benefit
  to traditional data mining to make better
  business decisions

10
Applications of Text Mining

• Understanding customers better
• Competitive Intelligence
• Automated document management
• Customer complaint analysis
• Cross-sell/Up-sell
• Survey analysis
• Enhanced data mining
• Quality control
• Customer support
• Government intelligence

• Research
• Product development
• Fraud analysis
• Public opinion
• Disease management
• Claims analysis
• Customer satisfaction
• Sales feedback analysis
• Email analysis for CRM
• Business performance monitoring

11
Information pipelines
Transactions
sales
12
Information pipelines
Transactions
sales
13
Text Mining Email Example
Inbound Email
Outbound Email
14
Taxonomy
15
Taxonomy
16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19
Text Mining
20
Text Mining
21
Text Mining
22
Text Mining
Email Response
23
Today's Agenda

• What is Text Mining ?
• Structured Text vs. Unstructured Text
• Linguistics and how it is used in Text Mining
• Applications of Text Mining
• How Text Mining is used through an email
  application to streamline a business process
• How Text Mining can be used to discover patterns
  and relationships in text to streamline a
  manufacturing business process

24
(No Transcript)
25
Warranty reports
First name
Last name
Street
Product
City
BMX 330 BMX 550 RBX 12 Sirax 220 Street E3
State
Zip
RBX 12
Replace
Repair
?
Problem description
26
Mining problem description fields
FX frame
chromoly butting
RBX 12 Mountain Bike
head tube
spokes
Fimano 33G derailleur
27
Mining problem description fields
FX frame
33
74
chromoly butting
RBX 12 Mountain Bike
head tube
spokes
Fimano 33G derailleur
28
Cost of a defect over time
cost
time
29
Cost of a defect over time
cost
time
30
Applications of Text Mining

• Understanding customers better
• Competitive Intelligence
• Automated document management
• Customer complaint analysis
• Cross-sell/Up-sell
• Survey analysis
• Enhanced data mining
• Quality control
• Customer support
• Government intelligence

• Research
• Product development
• Fraud analysis
• Public opinion
• Disease management
• Claims analysis
• Customer satisfaction
• Sales feedback analysis
• Email analysis for CRM
• Business performance monitoring

31
Business Intelligence

• Get the Right Information
• To the Right Person
• At the Right Time
• To Make the Right

Business Decisions
32
Conclusion and QA
Contact us or one of our other text mining
experts by calling 800-543-5815. David Purcell
dpurcell_at_spss.com Doug Bryan dbryan_at_spss.com