decision support and business intelligence systems...

Decision Support and Business Intelligence Systems

(9th Ed., Prentice Hall)

Chapter 7: Text and Web Mining

Copyright © 2011 Pearson Education, Inc. Publishing as Prentice Hall 7-2

Learning Objectives

n  Describe text mining and understand the need for text mining

n  Differentiate between text mining, Web mining and data mining

n  Understand the different application areas for text mining

n  Know the process of carrying out a text mining project

n  Understand the different methods to introduce structure to text-based data


Learning Objectives

n  Describe Web mining, its objectives, and its benefits

n  Understand the three different branches of Web mining n  Web content mining n  Web structure mining n  Web usage mining

n  Understand the applications of these three mining paradigms


Opening Vignette:

“Mining Text for Security and Counterterrorism”

n What is MITRE? n Problem description n Proposed solution n Results n Answer and discuss the case questions


Opening Vignette: Mining Text For Security…

(L) Kampala

(L) Uganda

(P) Yoweri Museveni

(L) Sudan

(L) Khartoum

(L) Southern Sudan

(P) Timothy McVeigh

(P) Oklahoma City

(P) Terry Nichols

(E) election

(P) Norodom Ranariddh

(P) Norodom Sihanouk

(L) Bangkok

(L) Cambodia

(L) Phnom Penh

(L) Thailand

(P) Hun Sen

(O) Khmer Rouge

(P) Pol Pot

Cluster 1 Cluster 2 Cluster 3


Text Mining Concepts n  85-90 percent of all corporate data is in some

kind of unstructured form (e.g., text) n  Unstructured corporate data is doubling in

size every 18 months n  Tapping into these information sources is not

an option, but a need to stay competitive n  Answer: text mining

n  A semi-automated process of extracting knowledge from unstructured data sources

n  a.k.a. text data mining or knowledge discovery in textual databases


Data Mining versus Text Mining

n  Both seek for novel and useful patterns n  Both are semi-automated processes n  Difference is the nature of the data:

n  Structured versus unstructured data n  Structured data: in databases n  Unstructured data: Word documents, PDF

files, text excerpts, XML files, and so on

n  Text mining – first, impose structure to the data, then mine the structured data


Text Mining Concepts n  Benefits of text mining are obvious especially

in text-rich data environments n  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.

n  Electronic communization records (e.g., Email) n  Spam filtering n  Email prioritization and categorization n  Automatic response generation


Text Mining Application Area

n  Information extraction n  Topic tracking n  Summarization n  Categorization n  Clustering n  Concept linking n  Question answering


Text Mining Terminology

n  Unstructured or semistructured data n  Corpus (and corpora) n  Terms n  Concepts n  Stemming n  Stop words (and include words) n  Synonyms (and polysemes) n  Tokenizing


Text Mining Terminology

n  Term dictionary n  Word frequency n  Part-of-speech tagging n  Morphology n  Term-by-document matrix

n  Occurrence matrix

n  Singular value decomposition n  Latent semantic indexing


Text Mining for Patent Analysis (see Applications Case 7.2)

n  What is a patent? n  “exclusive rights granted by a country to

an inventor for a limited period of time in exchange for a disclosure of an invention”

n  How do we do patent analysis (PA)? n  Why do we need to do PA?

n  What are the benefits? n  What are the challenges?

n  How does text mining help in PA?


Natural Language Processing (NLP)

n  Structuring a collection of text n  Old approach: bag-of-words n  New approach: natural language processing

n  NLP is … n  a very important concept in text mining n  a subfield of artificial intelligence and computational

linguistics n  the studies of "understanding" the natural human

language

n  Syntax versus semantics based text mining



n  What is “Understanding” ? n  Human understands, what about computers? n  Natural language is vague, context driven n  True understanding requires extensive knowledge

of a topic

n  Can/will computers ever understand natural language the same/accurate way we do?



n  Challenges in NLP n  Part-of-speech tagging n  Text segmentation n  Word sense disambiguation n  Syntax ambiguity n  Imperfect or irregular input n  Speech acts

n  Dream of AI community n  to have algorithms that are capable of automatically

reading and obtaining knowledge from text



n  WordNet n  A laboriously hand-coded database of English

words, their definitions, sets of synonyms, and various semantic relations between synonym sets

n  A major resource for NLP n  Need automation to be completed

n  Sentiment Analysis n  A technique used to detect favorable and

unfavorable opinions toward specific products and services

n  See Application Case 7.3 for a CRM application


NLP Task Categories

n  Information retrieval n  Information extraction n  Named-entity recognition n  Question answering n  Automatic summarization n  Natural language generation and understanding n  Machine translation n  Foreign language reading and writing n  Speech recognition n  Text proofing n  Optical character recognition


Text Mining Applications

n  Marketing applications n  Enables better CRM

n  Security applications n  ECHELON, OASIS n  Deception detection (…)

n  Medicine and biology n  Literature-based gene identification (…)

n  Academic applications n  Research stream analysis



n  Application Case 7.4: Mining for Lies n  Deception detection

n  A difficult problem n  If detection is limited to only text, then the

problem is even more difficult

n  The study n  analyzed text based testimonies of person

of interests at military bases n  used only text-based features (cues)



n  Application Case 7.4: Mining for Lies

Statements Transcribed for

Processing

Text Processing Software Identified Cues in Statements

Statements Labeled as Truthful or Deceptive By Law Enforcement

Text Processing Software Generated

Quantified Cues

Classification Models Trained and Tested on

Quantified Cues

Cues Extracted & Selected



n  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.



n  Application Case 7.4: Mining for Lies n  371 usable statements are generated n  31 features are used n  Different feature selection methods used n  10-fold cross validation is used n  Results (overall % accuracy)

n  Logistic regression 67.28 n  Decision trees 71.60 n  Neural networks 73.46


Text Mining Applications (gene/protein interaction identification)

Gen

e/P

rote

in 596 12043 24224 281020 42722 397276

D007962

D 016923

D 001773

D019254 D044465 D001769 D002477 D003643 D016158

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

NN IN NN IN VBZ IN JJ JJ NN NN NN CC NN IN NN

NP PP NP NP PP NP NP PP NP

Ont

olog

yW

ord

PO

SS

hallo

w

Par

se

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


Text Mining Process

Extract knowledge from available data sources

A0

Unstructured data (text)

Structured data (databases)

Context-specific knowledge

Software/hardware limitationsPrivacy issues

Tools and techniquesDomain expertise

Linguistic limitations

Context diagram for the text mining

process


Text Mining Process

Establish the Corpus:Collect & Organize the

Domain Specific Unstructured Data

Create the Term-Document Matrix:Introduce Structure

to the Corpus

Extract Knowledge:Discover Novel

Patterns from the T-D Matrix

The inputs to the process includes a variety of relevant unstructured (and semi-structured) data sources such as text, XML, HTML, etc.

The output of the Task 1 is a collection of documents in some digitized format for computer processing

The output of the Task 2 is a flat file called term-document matrix where the cells are populated with the term frequencies

The output of Task 3 is a number of problem specific classification, association, clustering models and visualizations

Task 1 Task 2 Task 3

FeedbackFeedback

The three-step text mining process


Text Mining Process

n  Step 1: Establish the corpus n  Collect all relevant unstructured data

(e.g., textual documents, XML files, emails, Web pages, short notes, voice recordings…)

n  Digitize, standardize the collection (e.g., all in ASCII text files)

n  Place the collection in a common place (e.g., in a flat file, or in a directory as separate files)


Text Mining Process

n  Step 2: Create the Term–by–Document Matrix

investment risk

project management

software engineering

development

1

SAP...

Document 1

Document 2

Document 3

Document 4

Document 5

Document 6

...

Documents

Terms

1

1

1

2

1

1

1

3

1


Text Mining Process

n  Step 2: Create the Term–by–Document Matrix (TDM), cont. n  Should all terms be included?

n  Stop words, include words n  Synonyms, homonyms n  Stemming

n  What is the best representation of the indices (values in cells)?

n  Row counts; binary frequencies; log frequencies; n  Inverse document frequency


Text Mining Process

n  Step 2: Create the Term–by–Document Matrix (TDM), cont. n  TDM is a sparse matrix. How can we reduce

the dimensionality of the TDM? n  Manual - a domain expert goes through it n  Eliminate terms with very few occurrences in

very few documents (?) n  Transform the matrix using singular value

decomposition (SVD) n  SVD is similar to principle component analysis


Text Mining Process

n  Step 3: Extract patterns/knowledge n  Classification (text categorization) n  Clustering (natural groupings of text)

n  Improve search recall n  Improve search precision n  Scatter/gather n  Query-specific clustering

n  Association n  Trend Analysis (…)


Text Mining Application (research trend identification in literature)

n  Mining the published IS literature n  MIS Quarterly (MISQ) n  Journal of MIS (JMIS) n  Information Systems Research (ISR)

n  Covers 12-year period (1994-2005) n  901 papers are included in the study n  Only the paper abstracts are used n  9 clusters are generated for further analysis



Journal Year Author(s) Title Vol/No Pages Keywords Abstract

MISQ 2005 A. Malhotra,S. Gosain andO. A. El Sawy

Absorptive capacity configurations in supply chains: Gearing for partner-enabled market knowledge creation

29/1 145-187 knowledge managementsupply chainabsorptive capacityinterorganizational information systemsconfiguration approaches

The need for continual value innovation is driving supply chains to evolve from a pure transactional focus to leveraging interorganizational partner ships for sharing

ISR 1999 D. Robey andM. C. Boudreau

Accounting for the contradictory organizational consequences of information technology: Theoretical directions and methodological implications

2-Oct 167-185 organizational transformationimpacts of technologyorganization theoryresearch methodologyintraorganizational powerelectronic communicationmis implementationculturesystems

Although much contemporary thought considers advanced information technologies as either determinants or enablers of radical organizational change, empirical studies have revealed inconsistent findings to support the deterministic logic implicit in such arguments. This paper reviews the contradictory

JMIS 2001 R. Aron andE. K. Clemons

Achieving the optimal balance between investment in quality and investment in self-promotion for information products

18/2 65-88 information productsinternet advertisingproduct positioningsignalingsignaling games

When producers of goods (or services) are confronted by a situation in which their offerings no longer perfectly match consumer preferences, they must determine the extent to which the advertised features of

… … … … … … … …



YEAR

No

of A

rticl

es

CLUSTER: 1

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

05

101520253035

CLUSTER: 2

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

CLUSTER: 3

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

CLUSTER: 4

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

05

101520253035

CLUSTER: 5

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

CLUSTER: 6

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

CLUSTER: 7

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

05

101520253035

CLUSTER: 8

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

CLUSTER: 9

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005



JOURNAL

No

of A

rticl

es

CLUSTER: 1

ISR JMIS MISQ0

102030405060708090

100

CLUSTER: 2

ISR JMIS MISQ

CLUSTER: 3

ISR JMIS MISQ

CLUSTER: 4

ISR JMIS MISQ0

102030405060708090

100

CLUSTER: 5

ISR JMIS MISQ

CLUSTER: 6

ISR JMIS MISQ

CLUSTER: 7

ISR JMIS MISQ0

102030405060708090

100

CLUSTER: 8

ISR JMIS MISQ

CLUSTER: 9

ISR JMIS MISQ


Text Mining Tools

n  Commercial Software Tools n  SPSS PASW Text Miner n  SAS Enterprise Miner n  Statistica Data Miner n  ClearForest, …

n  Free Software Tools n  RapidMiner n  GATE n  Spy-EM, …


Web Mining Overview

n  Web is the largest repository of data n  Data is in HTML, XML, text format n  Challenges (of processing Web data)

n  The Web is too big for effective data mining n  The Web is too complex n  The Web is too dynamic n  The Web is not specific to a domain n  The Web has everything

n  Opportunities and challenges are great!


Web Mining

n  Web mining (or Web data mining) is the process of discovering intrinsic relationships from Web data (textual, linkage, or usage)

Web Mining

Web Structure MiningSource: the unified

resource locator (URL) links contained in the

Web pages

Web Content MiningSource: unstructured textual content of the Web pages (usually in

HTML format)

Web Usage MiningSource: the detailed description of a Web

site’s visits (sequence of clicks by sessions)


Web Content/Structure Mining

n  Mining of the textual content on the Web n  Data collection via Web crawlers

n  Web pages include hyperlinks n  Authoritative pages n  Hubs n  hyperlink-induced topic search (HITS) alg


Web Usage Mining

n  Extraction of information from data generated through Web page visits and transactions… n  data stored in server access logs, referrer logs,

agent logs, and client-side cookies n  user characteristics and usage profiles n  metadata, such as page attributes, content

attributes, and usage data

n  Clickstream data n  Clickstream analysis


Web Usage Mining

n  Web usage mining applications n  Determine the lifetime value of clients n  Design cross-marketing strategies across products. n  Evaluate promotional campaigns n  Target electronic ads and coupons at user groups

based on user access patterns n  Predict user behavior based on previously learned

rules and users' profiles n  Present dynamic information to users based on

their interests and profiles…


Web Usage Mining (clickstream analysis)

Weblogs

Website Pre-Process Data Collecting Merging Cleaning Structuring - Identify users - Identify sessions - Identify page views - Identify visits

Extract Knowledge Usage patterns User profiles Page profiles Visit profiles Customer value

How to better the data

How to improve the Web site

How to increase the customer value

User /Customer


Web Mining Success Stories

n  Amazon.com, Ask.com, Scholastic.com, … n  Website Optimization Ecosystem

Web Analytics

Voice of Customer

Customer Experience Management

Customer Interaction on the Web

Analysis of Interactions Knowledge about the Holistic View of the Customer


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


End of the Chapter

n  Questions / comments…


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