Decision Support and Business Intelligence Systems
(9
thEd., Prentice Hall) Chapter 7:
Text and Web Mining
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
Natural Language Processing (NLP)
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?
Natural Language Processing (NLP)
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
Natural Language Processing (NLP)
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
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
Research stream analysis
Text Mining Applications
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
used only text-based features (cues)
Text Mining Applications
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
Text Mining Applications
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 …
Text Mining Applications
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)
Gene/ Protein 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
OntologyWordPOShallow arse
...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 limitations
Privacy issues
Tools and techniques Domain 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
Feedback Feedback
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
Text Mining Application
(research trend identification in literature)
Journal Year Author(s) Title Vol/No Pages Keywords Abstract
MISQ 2005 A. Malhotra, S. Gosain and O. A. El Sawy
Absorptive capacity configurations in supply chains:
Gearing for partner- enabled market knowledge creation
29/1 145-187 knowledge management supply chain
absorptive capacity interorganizational information systems configuration 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 and
M. C. Boudreau
Accounting for the contradictory organizational consequences of information technology:
Theoretical directions and methodological implications
2-Oct 167-185 organizational transformation
impacts of technology organization theory research methodology intraorganizational power electronic communication mis implementation culture
systems
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 and
E. K. Clemons
Achieving the optimal balance between investment in quality and investment in self- promotion for
information products
18/2 65-88 information products internet advertising product positioning signaling
signaling 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
Text Mining Application
(research trend identification in literature)
No of Articles
CLUSTER: 1
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
0 5 10 15 20 25 30 35
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
0 5 10 15 20 25 30 35
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
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
0 5 10 15 20 25 30 35
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Text Mining Application
(research trend identification in literature)
JOURNAL
No of Articles
CLUSTER: 1 ISR JMIS MISQ 0
10 20 30 40 50 60 70 80 90 100
CLUSTER: 2 ISR JMIS MISQ
CLUSTER: 3 ISR JMIS MISQ
CLUSTER: 4 ISR JMIS MISQ 0
10 20 30 40 50 60 70 80 90 100
CLUSTER: 5 ISR JMIS MISQ
CLUSTER: 6 ISR JMIS MISQ
CLUSTER: 7 ISR JMIS MISQ 0
10 20 30 40 50 60 70 80 90 100
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
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 Mining Source: the unified resource locator (URL)
links contained in the Web pages Web Content Mining
Source: unstructured textual content of the Web pages (usually in
HTML format)
Web Usage Mining Source: 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
