Decision Support and Business

Intelligence Systems
(9th Ed., Prentice Hall)

Chapter 7:
Text and Web Mining
 Learning Objectives
 Describe text mining and understand the need
for text mining
 Differentiate between text mining, Web mining
and data mining
 Understand the different application areas for
text mining
 Know the process of carrying out a text mining
project
 Understand the different methods to introduce
structure to text-based data
-2 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 Learning Objectives
 Describe Web mining, its objectives, and its
benefits
 Understand the three different branches of
Web mining
 Web content mining
 Web structure mining
 Web usage mining
 Understand the applications of these three
mining paradigms

-3 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Opening Vignette:
“Mining Text for Security and
Counterterrorism”
What is MITRE?

Problem description

Proposed solution

Results

Answer and discuss the case questions

-4 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Opening Vignette:
Mining Text For Security…

Cluster 1 Cluster 2 Cluster 3

(L) Kampala (P) Timothy McVeigh (E) election
(L) Uganda (P) Oklahoma City (P) Norodom Ranariddh
(P) Yoweri Museveni (P) Terry Nichols (P) Norodom Sihanouk
(L) Sudan (L) Bangkok
(L) Khartoum (L) Cambodia
(L) Southern Sudan (L) Phnom Penh
(L) Thailand
(P) Hun Sen
(O) Khmer Rouge
(P) Pol Pot

-5 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Text Mining Concepts
 85-90 percent of all corporate data is in some
kind of unstructured form (e.g., text)
 Unstructured corporate data is doubling in size
every 18 months
 Tapping into these information sources is not an
option, but a need to stay competitive
 Answer: text mining
 A semi-automated process of extracting knowledge
from unstructured data sources
 a.k.a. text data mining or knowledge discovery in
textual databases
-6 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 Data Mining versus Text Mining
 Both seek for novel and useful patterns
 Both are semi-automated processes
 Difference is the nature of the data:
 Structured versus unstructured data
 Structured data: in databases
 Unstructured data: Word documents, PDF
files, text excerpts, XML files, and so on
 Text mining – first, impose structure to
the data, then mine the structured data
-7 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 Text Mining Concepts
 Benefits of text mining are obvious especially
in text-rich data environments
 e.g., law (court orders), academic research
(research articles), finance (quarterly reports),
medicine (discharge summaries), biology (molecular
interactions), technology (patent files), marketing
(customer comments), etc.
 Electronic communization records (e.g., Email)
 Spam filtering
 Email prioritization and categorization
 Automatic response generation
-8 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 Text Mining Application Area
 Information extraction
 Topic tracking
 Summarization
 Categorization
 Clustering
 Concept linking
 Question answering

-9 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Text Mining Terminology
 Unstructured or semistructured data
 Corpus (and corpora)
 Terms
 Concepts
 Stemming
 Stop words (and include words)
 Synonyms (and polysemes)
 Tokenizing
-10 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 Text Mining Terminology
 Term dictionary
 Word frequency
 Part-of-speech tagging
 Morphology
 Term-by-document matrix
 Occurrence matrix
 Singular value decomposition
 Latent semantic indexing
-11 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 Text Mining for Patent Analysis
(see Applications Case 7.2)
 What is a patent?
 “exclusive rights granted by a country to
an inventor for a limited period of time in
exchange for a disclosure of an invention”
 How do we do patent analysis (PA)?
 Why do we need to do PA?
 What are the benefits?
 What are the challenges?
 How does text mining help in PA?
-12 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 Natural Language Processing (NLP)
 Structuring a collection of text
 Old approach: bag-of-words
 New approach: natural language processing
 NLP is …
 a very important concept in text mining
 a subfield of artificial intelligence and computational
linguistics
 the studies of "understanding" the natural human
language
 Syntax versus semantics based text mining
-13 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 Natural Language Processing (NLP)
 What is “Understanding” ?
 Human understands, what about computers?
 Natural language is vague, context driven
 True understanding requires extensive knowledge
of a topic

 Can/will computers ever understand natural

language the same/accurate way we do?

-14 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Natural Language Processing (NLP)
 Challenges in NLP
 Part-of-speech tagging
 Text segmentation
 Word sense disambiguation
 Syntax ambiguity
 Imperfect or irregular input
 Speech acts

 Dream of AI community
 to have algorithms that are capable of automatically
reading and obtaining knowledge from text
-15 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 Natural Language Processing (NLP)
 WordNet
 A laboriously hand-coded database of English
words, their definitions, sets of synonyms, and
various semantic relations between synonym sets
 A major resource for NLP
 Need automation to be completed
 Sentiment Analysis
 A technique used to detect favorable and
unfavorable opinions toward specific products and
services
 See Application Case 7.3 for a CRM application
-16 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 NLP Task Categories
 Information retrieval
 Information extraction
 Named-entity recognition
 Question answering
 Automatic summarization
 Natural language generation and understanding
 Machine translation
 Foreign language reading and writing
 Speech recognition
 Text proofing
 Optical character recognition
-17 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 Text Mining Applications
 Marketing applications
 Enables better CRM
 Security applications
 ECHELON, OASIS
 Deception detection (…)
 Medicine and biology
 Literature-based gene identification (…)
 Academic applications
 Research stream analysis
-18 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 Text Mining Applications
 Application Case 7.4: Mining for Lies
 Deception detection
 A difficult problem
 If detection is limited to only text, then the
problem is even more difficult
 The study
 analyzed text based testimonies of person
of interests at military bases
 used only text-based features (cues)
-19 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 Text Mining Applications
 Application Case 7.4: Mining for Lies
Statements
Transcribed for
Processing

Statements Labeled as
Cues Extracted &
Truthful or Deceptive
Selected
By Law Enforcement

Classification Models Text Processing

Trained and Tested on Software Identified
Quantified Cues Cues in Statements

Text Processing
Software Generated
Quantified Cues

-20 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Text Mining Applications
 Application Case 7.4: Mining for Lies
Category Example Cues
Quantity Verb count, noun-phrase count, ...
Complexity Avg. no of clauses, sentence length, …
Uncertainty Modifiers, modal verbs, ...
Nonimmediacy Passive voice, objectification, ...
Expressivity Emotiveness
Diversity Lexical diversity, redundancy, ...
Informality Typographical error ratio
Specificity Spatiotemporal, perceptual information …
Affect Positive affect, negative affect, etc.

-21 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Text Mining Applications
 Application Case 7.4: Mining for Lies
 371 usable statements are generated
 31 features are used
 Different feature selection methods used
 10-fold cross validation is used
 Results (overall % accuracy)
 Logistic regression 67.28
 Decision trees 71.60
 Neural networks 73.46

-22 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Text Mining Applications
(gene/protein interaction identification)
Protein

596 12043 24224 281020 42722 397276

Gene/

D007962

D 016923
Ontology

D 001773

D019254 D044465 D001769 D002477 D003643 D016158

...expression of Bcl-2 is correlated with insufficient white blood cell death and activation of p53.
Word

185 8 51112 9 23017 27 5874 2791 8952 1623 5632 17 8252 8 2523
POS

NN IN NN IN VBZ IN JJ JJ NN NN NN CC NN IN NN
Shallow
Parse

NP PP NP NP PP NP NP PP NP

-23 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Text Mining Process

Context diagram for

the text mining Software/hardware
Privacy issues
process Linguistic limit
Extract
Context-specific
Unstructured data
knowledge
(text) knowledge
from available

Structured data (databases)

data sources

A0
Domain expertis
Tools and techniqu
-24 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 Text Mining Process

Task 1 Task 2 Task 3

Establish the Corpus: Create the Term- Extract Knowledge:

Collect & Organize the Document Matrix: Discover Novel
Domain Specific Introduce Structure Patterns from the
Unstructured Data to the Corpus T-D Matrix

Feedback Feedback

The inputs to the process The output of the Task 1 is a The output of the Task 2 is a The output of Task 3 is a
includes a variety of relevant collection of documents in flat file called term-document number of problem specific
unstructured (and semi- some digitized format for matrix where the cells are classification, association,
structured) data sources such computer processing populated with the term clustering models and
as text, XML, HTML, etc. frequencies visualizations

The three-step text mining process

-25 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Text Mining Process
 Step 1: Establish the corpus
 Collect all relevant unstructured data
(e.g., textual documents, XML files, emails,
Web pages, short notes, voice recordings…)
 Digitize, standardize the collection
(e.g., all in ASCII text files)
 Place the collection in a common place
(e.g., in a flat file, or in a directory as
separate files)

-26 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Text Mining Process
 Step 2: Create the Term–by–Document Matrix
e nt
er ing
Terms m ine
r i sk age g t
ent an
e e n
m en
m ar
e stm j e ct f tw e lo p P
v
Documents inv pro so de SA ...
Document 1 1 1

Document 2 1

Document 3 3 1

Document 4 1

Document 5 2 1

Document 6 1 1
...

-27 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Text Mining Process
 Step 2: Create the Term–by–Document
Matrix (TDM), cont.
 Should all terms be included?
 Stop words, include words
 Synonyms, homonyms
 Stemming
 What is the best representation of the
indices (values in cells)?
 Row counts; binary frequencies; log frequencies;
 Inverse document frequency
-28 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 Text Mining Process
 Step 2: Create the Term–by–Document
Matrix (TDM), cont.
 TDM is a sparse matrix. How can we reduce
the dimensionality of the TDM?
 Manual - a domain expert goes through it
 Eliminate terms with very few occurrences in
very few documents (?)
 Transform the matrix using singular value
decomposition (SVD)
 SVD is similar to principle component analysis

-29 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Text Mining Process
 Step 3: Extract patterns/knowledge
 Classification (text categorization)
 Clustering (natural groupings of text)
 Improve search recall
 Improve search precision
 Scatter/gather
 Query-specific clustering
 Association
 Trend Analysis (…)

-30 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Text Mining Application
(research trend identification in literature)
 Mining the published IS literature
 MIS Quarterly (MISQ)
 Journal of MIS (JMIS)
 Information Systems Research (ISR)

 Covers 12-year period (1994-2005)

 901 papers are included in the study
 Only the paper abstracts are used
 9 clusters are generated for further analysis
-31 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 Text Mining Application
(research trend identification in literature)
Journal Year Author(s) Title Vol/No Pages Keywords Abstract
MISQ 2005 A. Malhotra, Absorptive capacity 29/1 145-187 knowledge management The need for continual value
S. Gosain and configurations in supply chain innovation is driving supply
O. A. El Sawy supply chains: absorptive capacity chains to evolve from a pure
Gearing for partner- interorganizational transactional focus to
enabled market information systems leveraging interorganizational
knowledge creation configuration approaches partner ships for sharing
ISR 1999 D. Robey and Accounting for the 2-Oct 167-185 organizational Although much contemporary
M. C. Boudreau contradictory transformation thought considers advanced
organizational impacts of technology information technologies as
consequences of organization theory either determinants or enablers
information research methodology of radical organizational
technology: intraorganizational power change, empirical studies have
Theoretical directions electronic communication revealed inconsistent findings to
and methodological mis implementation support the deterministic logic
implications culture implicit in such arguments. This
systems paper reviews the contradictory
JMIS 2001 R. Aron and Achieving the optimal 18/2 65-88 information products When producers of goods (or
E. K. Clemons balance between internet advertising services) are confronted by a
investment in quality product positioning situation in which their offerings
and investment in self- signaling no longer perfectly match
promotion for signaling games consumer preferences, they
information products must determine the extent to
which the advertised features of

… … … … … … … …

-32 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

-33
No of Articles

1
1
2
2
3
3
1
1
2
2
3
3
1
1
2
2
3
3

0
5
0
5
0
5
0
5
0
5
0
5
0
5
0
5
0
5

0
5
0
5
0
5
1994 1994 1994
1995 1995 1995
1996 1996 1996
1997 1997 1997
1998 1998 1998
1999 1999 1999
2000 2000 2000
2001 2001 2001

C LU S TER : 7
C LU S TER : 4
C LU S TER : 1

2002 2002 2002

2003 2003 2003
2004 2004 2004
2005 2005 2005

1994 1994 1994

1995 1995 1995
1996 1996 1996
1997 1997 1997
1998 1998 1998
1999 1999 1999

Y EAR
2000 2000 2000
2001 2001 2001

C LU STER : 8
C LU STER : 5
C LU STER : 2

2002 2002 2002

2003 2003 2003
2004 2004 2004
2005 2005 2005
Text Mining Application

Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

1994 1994 1994
1995 1995 1995
1996 1996 1996
1997 1997 1997
1998 1998 1998
1999 1999 1999
2000 2000 2000
2001 2001 2001
C LU STER : 9
C LU STER : 6
C LU STER : 3

2002 2002 2002

2003 2003 2003
2004 2004 2004
2005 2005 2005
(research trend identification in literature)
 Text Mining Application
(research trend identification in literature)
100
90
80
70
60
50
40
30
20
10
0
IS R J M IS M IS Q IS R J M IS M IS Q IS R J M IS M IS Q

C LU S T ER : 1 C LU S T ER : 2 C LU S T E R : 3
100
90
80
70
No of Articles

60
50
40
30
20
10
0
IS R J M IS M IS Q IS R J M IS M IS Q IS R J M IS M IS Q

C LU S T ER : 4 C LU S T ER : 5 C LU S T E R : 6
100
90
80
70
60
50
40
30
20
10
0
IS R J M IS M IS Q IS R J M IS M IS Q IS R J M IS M IS Q

C LU S T ER : 7 C LU S T ER : 8 C LU S T E R : 9

JO U R N AL

-34 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Text Mining Tools
 Commercial Software Tools
 SPSS PASW Text Miner
 SAS Enterprise Miner
 Statistica Data Miner
 ClearForest, …
 Free Software Tools
 RapidMiner
 GATE
 Spy-EM, …
-35 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 Web Mining Overview
 Web is the largest repository of data
 Data is in HTML, XML, text format
 Challenges (of processing Web data)
 The Web is too big for effective data mining
 The Web is too complex
 The Web is too dynamic
 The Web is not specific to a domain
 The Web has everything

 Opportunities and challenges are great!

-36 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall
 Web Mining
 Web mining (or Web data mining) is the
process of discovering intrinsic relationships
from Web data (textual, linkage, or usage)

Web Mining

Web Content Mining Web Structure Mining Web Usage Mining

Source: unstructured Source: the unified Source: the detailed
textual content of the resource locator (URL) description of a Web
Web pages (usually in links contained in the site’s visits (sequence
HTML format) Web pages of clicks by sessions)

-37 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Web Content/Structure Mining
 Mining of the textual content on the Web
 Data collection via Web crawlers

 Web pages include hyperlinks

 Authoritative pages
 Hubs
 hyperlink-induced topic search (HITS) alg

-38 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Web Usage Mining
 Extraction of information from data generated
through Web page visits and transactions…
 data stored in server access logs, referrer logs,
agent logs, and client-side cookies
 user characteristics and usage profiles
 metadata, such as page attributes, content
attributes, and usage data
 Clickstream data
 Clickstream analysis

-39 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Web Usage Mining
 Web usage mining applications
 Determine the lifetime value of clients
 Design cross-marketing strategies across products.
 Evaluate promotional campaigns
 Target electronic ads and coupons at user groups
based on user access patterns
 Predict user behavior based on previously learned
rules and users' profiles
 Present dynamic information to users based on
their interests and profiles…

-40 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Web Usage Mining
(clickstream analysis)

Pre-Process Data Extract Knowledge

Website
User / Collecting Usage patterns
Customer Merging User profiles
Cleaning Page profiles
Structuring Visit profiles
- Identify users Customer value
- Identify sessions
- Identify page views
- Identify visits
Weblogs

How to better the data

How to improve the Web site
How to increase the customer value

-41 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Web Mining Success Stories
 Amazon.com, Ask.com, Scholastic.com, …
 Website Optimization Ecosystem
Customer Interaction Analysis of Interactions Knowledge about the Holistic
on the Web View of the Customer

Web
Analytics

Voice of
Customer

Customer Experience
Management

-42 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 Web Mining Tools
Product Name URL
Angoss Knowledge WebMiner angoss.com
ClickTracks clicktracks.com
LiveStats from DeepMetrix deepmetrix.com
Megaputer WebAnalyst megaputer.com
MicroStrategy Web Traffic Analysis microstrategy.com
SAS Web Analytics sas.com
SPSS Web Mining for Clementine spss.com
WebTrends webtrends.com
XML Miner scientio.com

-43 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 End of the Chapter

 Questions / comments…

-44 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall

 All rights reserved. No part of this publication may be reproduced, stored in a
retrieval system, or transmitted, in any form or by any means, electronic,
mechanical, photocopying, recording, or otherwise, without the prior written
permission of the publisher. Printed in the United States of America.

Copyright © 2011 Pearson Education, Inc.  

Publishing as Prentice Hall

-45 Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall