Jiawei Han, Micheline Kamber and Jian Pei
Data Mining: Concepts and Techniques, 3rd ed.
The Morgan Kaufmann Series in Data Management Systems
Morgan Kaufmann Publishers, July 2011. ISBN 978-0123814791
1. 1
1
Data Mining:
Concepts and Techniques
(3rd ed.)
Chapter 1
Subrata Kumer Paul
Assistant Professor, Dept. of CSE, BAUET
sksubrata96@gmail.com
2. 2
Chapter 1. Introduction
īŽ Why Data Mining?
īŽ What Is Data Mining?
īŽ A Multi-Dimensional View of Data Mining
īŽ What Kind of Data Can Be Mined?
īŽ What Kinds of Patterns Can Be Mined?
īŽ What Technology Are Used?
īŽ What Kind of Applications Are Targeted?
īŽ Major Issues in Data Mining
īŽ A Brief History of Data Mining and Data Mining Society
īŽ Summary
3. 3
Why Data Mining?
īŽ The Explosive Growth of Data: from terabytes to petabytes
īŽ Data collection and data availability
īŽ Automated data collection tools, database systems, Web,
computerized society
īŽ Major sources of abundant data
īŽ Business: Web, e-commerce, transactions, stocks, âĻ
īŽ Science: Remote sensing, bioinformatics, scientific simulation, âĻ
īŽ Society and everyone: news, digital cameras, YouTube
īŽ We are drowning in data, but starving for knowledge!
īŽ âNecessity is the mother of inventionââData miningâAutomated
analysis of massive data sets
4. 4
Evolution of Sciences
īŽ Before 1600, empirical science
īŽ 1600-1950s, theoretical science
īŽ Each discipline has grown a theoretical component. Theoretical models often
motivate experiments and generalize our understanding.
īŽ 1950s-1990s, computational science
īŽ Over the last 50 years, most disciplines have grown a third, computational branch
(e.g. empirical, theoretical, and computational ecology, or physics, or linguistics.)
īŽ Computational Science traditionally meant simulation. It grew out of our inability to
find closed-form solutions for complex mathematical models.
īŽ 1990-now, data science
īŽ The flood of data from new scientific instruments and simulations
īŽ The ability to economically store and manage petabytes of data online
īŽ The Internet and computing Grid that makes all these archives universally accessible
īŽ Scientific info. management, acquisition, organization, query, and visualization tasks
scale almost linearly with data volumes. Data mining is a major new challenge!
īŽ Jim Gray and Alex Szalay, The World Wide Telescope: An Archetype for Online Science,
Comm. ACM, 45(11): 50-54, Nov. 2002
5. 5
Evolution of Database Technology
īŽ 1960s:
īŽ Data collection, database creation, IMS and network DBMS
īŽ 1970s:
īŽ Relational data model, relational DBMS implementation
īŽ 1980s:
īŽ RDBMS, advanced data models (extended-relational, OO, deductive, etc.)
īŽ Application-oriented DBMS (spatial, scientific, engineering, etc.)
īŽ 1990s:
īŽ Data mining, data warehousing, multimedia databases, and Web
databases
īŽ 2000s
īŽ Stream data management and mining
īŽ Data mining and its applications
īŽ Web technology (XML, data integration) and global information systems
6. 6
Chapter 1. Introduction
īŽ Why Data Mining?
īŽ What Is Data Mining?
īŽ A Multi-Dimensional View of Data Mining
īŽ What Kind of Data Can Be Mined?
īŽ What Kinds of Patterns Can Be Mined?
īŽ What Technology Are Used?
īŽ What Kind of Applications Are Targeted?
īŽ Major Issues in Data Mining
īŽ A Brief History of Data Mining and Data Mining Society
īŽ Summary
7. 7
What Is Data Mining?
īŽ Data mining (knowledge discovery from data)
īŽ Extraction of interesting (non-trivial, implicit, previously
unknown and potentially useful) patterns or knowledge from
huge amount of data
īŽ Data mining: a misnomer?
īŽ Alternative names
īŽ Knowledge discovery (mining) in databases (KDD), knowledge
extraction, data/pattern analysis, data archeology, data
dredging, information harvesting, business intelligence, etc.
īŽ Watch out: Is everything âdata miningâ?
īŽ Simple search and query processing
īŽ (Deductive) expert systems
8. 8
Knowledge Discovery (KDD) Process
īŽ This is a view from typical
database systems and data
warehousing communities
īŽ Data mining plays an essential
role in the knowledge discovery
process
Data Cleaning
Data Integration
Databases
Data Warehouse
Task-relevant Data
Selection
Data Mining
Pattern Evaluation
9. 9
Example: A Web Mining Framework
īŽ Web mining usually involves
īŽ Data cleaning
īŽ Data integration from multiple sources
īŽ Warehousing the data
īŽ Data cube construction
īŽ Data selection for data mining
īŽ Data mining
īŽ Presentation of the mining results
īŽ Patterns and knowledge to be used or stored into
knowledge-base
10. 10
Data Mining in Business Intelligence
Increasing potential
to support
business decisions End User
Business
Analyst
Data
Analyst
DBA
Decision
Making
Data Presentation
Visualization Techniques
Data Mining
Information Discovery
Data Exploration
Statistical Summary, Querying, and Reporting
Data Preprocessing/Integration, Data Warehouses
Data Sources
Paper, Files, Web documents, Scientific experiments, Database Systems
11. 11
Example: Mining vs. Data Exploration
īŽ Business intelligence view
īŽ Warehouse, data cube, reporting but not much mining
īŽ Business objects vs. data mining tools
īŽ Supply chain example: tools
īŽ Data presentation
īŽ Exploration
12. 12
KDD Process: A Typical View from ML and
Statistics
Input Data Data
Mining
Data Pre-
Processing
Post-
Processing
īŽ This is a view from typical machine learning and statistics communities
Data integration
Normalization
Feature selection
Dimension reduction
Pattern discovery
Association & correlation
Classification
Clustering
Outlier analysis
âĻ âĻ âĻ âĻ
Pattern evaluation
Pattern selection
Pattern interpretation
Pattern visualization
13. 13
Example: Medical Data Mining
īŽ Health care & medical data mining â often
adopted such a view in statistics and machine
learning
īŽ Preprocessing of the data (including feature
extraction and dimension reduction)
īŽ Classification or/and clustering processes
īŽ Post-processing for presentation
14. 14
Chapter 1. Introduction
īŽ Why Data Mining?
īŽ What Is Data Mining?
īŽ A Multi-Dimensional View of Data Mining
īŽ What Kind of Data Can Be Mined?
īŽ What Kinds of Patterns Can Be Mined?
īŽ What Technology Are Used?
īŽ What Kind of Applications Are Targeted?
īŽ Major Issues in Data Mining
īŽ A Brief History of Data Mining and Data Mining Society
īŽ Summary
15. 15
Multi-Dimensional View of Data Mining
īŽ Data to be mined
īŽ Database data (extended-relational, object-oriented, heterogeneous,
legacy), data warehouse, transactional data, stream, spatiotemporal,
time-series, sequence, text and web, multi-media, graphs & social
and information networks
īŽ Knowledge to be mined (or: Data mining functions)
īŽ Characterization, discrimination, association, classification,
clustering, trend/deviation, outlier analysis, etc.
īŽ Descriptive vs. predictive data mining
īŽ Multiple/integrated functions and mining at multiple levels
īŽ Techniques utilized
īŽ Data-intensive, data warehouse (OLAP), machine learning, statistics,
pattern recognition, visualization, high-performance, etc.
īŽ Applications adapted
īŽ Retail, telecommunication, banking, fraud analysis, bio-data mining,
stock market analysis, text mining, Web mining, etc.
16. 16
Chapter 1. Introduction
īŽ Why Data Mining?
īŽ What Is Data Mining?
īŽ A Multi-Dimensional View of Data Mining
īŽ What Kind of Data Can Be Mined?
īŽ What Kinds of Patterns Can Be Mined?
īŽ What Technology Are Used?
īŽ What Kind of Applications Are Targeted?
īŽ Major Issues in Data Mining
īŽ A Brief History of Data Mining and Data Mining Society
īŽ Summary
17. 17
Data Mining: On What Kinds of Data?
īŽ Database-oriented data sets and applications
īŽ Relational database, data warehouse, transactional database
īŽ Advanced data sets and advanced applications
īŽ Data streams and sensor data
īŽ Time-series data, temporal data, sequence data (incl. bio-sequences)
īŽ Structure data, graphs, social networks and multi-linked data
īŽ Object-relational databases
īŽ Heterogeneous databases and legacy databases
īŽ Spatial data and spatiotemporal data
īŽ Multimedia database
īŽ Text databases
īŽ The World-Wide Web
18. 18
Chapter 1. Introduction
īŽ Why Data Mining?
īŽ What Is Data Mining?
īŽ A Multi-Dimensional View of Data Mining
īŽ What Kind of Data Can Be Mined?
īŽ What Kinds of Patterns Can Be Mined?
īŽ What Technology Are Used?
īŽ What Kind of Applications Are Targeted?
īŽ Major Issues in Data Mining
īŽ A Brief History of Data Mining and Data Mining Society
īŽ Summary
19. 19
Data Mining Function: (1) Generalization
īŽ Information integration and data warehouse construction
īŽ Data cleaning, transformation, integration, and
multidimensional data model
īŽ Data cube technology
īŽ Scalable methods for computing (i.e., materializing)
multidimensional aggregates
īŽ OLAP (online analytical processing)
īŽ Multidimensional concept description: Characterization
and discrimination
īŽ Generalize, summarize, and contrast data
characteristics, e.g., dry vs. wet region
20. 20
Data Mining Function: (2) Association and
Correlation Analysis
īŽ Frequent patterns (or frequent itemsets)
īŽ What items are frequently purchased together in your
Walmart?
īŽ Association, correlation vs. causality
īŽ A typical association rule
īŽ Diaper ī Beer [0.5%, 75%] (support, confidence)
īŽ Are strongly associated items also strongly correlated?
īŽ How to mine such patterns and rules efficiently in large
datasets?
īŽ How to use such patterns for classification, clustering,
and other applications?
21. 21
Data Mining Function: (3) Classification
īŽ Classification and label prediction
īŽ Construct models (functions) based on some training examples
īŽ Describe and distinguish classes or concepts for future prediction
īŽ E.g., classify countries based on (climate), or classify cars
based on (gas mileage)
īŽ Predict some unknown class labels
īŽ Typical methods
īŽ Decision trees, naïve Bayesian classification, support vector
machines, neural networks, rule-based classification, pattern-
based classification, logistic regression, âĻ
īŽ Typical applications:
īŽ Credit card fraud detection, direct marketing, classifying stars,
diseases, web-pages, âĻ
22. 22
Data Mining Function: (4) Cluster Analysis
īŽ Unsupervised learning (i.e., Class label is unknown)
īŽ Group data to form new categories (i.e., clusters), e.g.,
cluster houses to find distribution patterns
īŽ Principle: Maximizing intra-class similarity & minimizing
interclass similarity
īŽ Many methods and applications
23. 23
Data Mining Function: (5) Outlier Analysis
īŽ Outlier analysis
īŽ Outlier: A data object that does not comply with the general
behavior of the data
īŽ Noise or exception? â One personâs garbage could be another
personâs treasure
īŽ Methods: by product of clustering or regression analysis, âĻ
īŽ Useful in fraud detection, rare events analysis
24. 24
Time and Ordering: Sequential Pattern,
Trend and Evolution Analysis
īŽ Sequence, trend and evolution analysis
īŽ Trend, time-series, and deviation analysis: e.g.,
regression and value prediction
īŽ Sequential pattern mining
īŽ e.g., first buy digital camera, then buy large SD
memory cards
īŽ Periodicity analysis
īŽ Motifs and biological sequence analysis
īŽ Approximate and consecutive motifs
īŽ Similarity-based analysis
īŽ Mining data streams
īŽ Ordered, time-varying, potentially infinite, data streams
25. 25
Structure and Network Analysis
īŽ Graph mining
īŽ Finding frequent subgraphs (e.g., chemical compounds), trees
(XML), substructures (web fragments)
īŽ Information network analysis
īŽ Social networks: actors (objects, nodes) and relationships (edges)
īŽ e.g., author networks in CS, terrorist networks
īŽ Multiple heterogeneous networks
īŽ A person could be multiple information networks: friends,
family, classmates, âĻ
īŽ Links carry a lot of semantic information: Link mining
īŽ Web mining
īŽ Web is a big information network: from PageRank to Google
īŽ Analysis of Web information networks
īŽ Web community discovery, opinion mining, usage mining, âĻ
26. 26
Evaluation of Knowledge
īŽ Are all mined knowledge interesting?
īŽ One can mine tremendous amount of âpatternsâ and knowledge
īŽ Some may fit only certain dimension space (time, location, âĻ)
īŽ Some may not be representative, may be transient, âĻ
īŽ Evaluation of mined knowledge â directly mine only
interesting knowledge?
īŽ Descriptive vs. predictive
īŽ Coverage
īŽ Typicality vs. novelty
īŽ Accuracy
īŽ Timeliness
īŽ âĻ
27. 27
Chapter 1. Introduction
īŽ Why Data Mining?
īŽ What Is Data Mining?
īŽ A Multi-Dimensional View of Data Mining
īŽ What Kind of Data Can Be Mined?
īŽ What Kinds of Patterns Can Be Mined?
īŽ What Technology Are Used?
īŽ What Kind of Applications Are Targeted?
īŽ Major Issues in Data Mining
īŽ A Brief History of Data Mining and Data Mining Society
īŽ Summary
28. 28
Data Mining: Confluence of Multiple Disciplines
Data Mining
Machine
Learning
Statistics
Applications
Algorithm
Pattern
Recognition
High-Performance
Computing
Visualization
Database
Technology
29. 29
Why Confluence of Multiple Disciplines?
īŽ Tremendous amount of data
īŽ Algorithms must be highly scalable to handle such as tera-bytes of
data
īŽ High-dimensionality of data
īŽ Micro-array may have tens of thousands of dimensions
īŽ High complexity of data
īŽ Data streams and sensor data
īŽ Time-series data, temporal data, sequence data
īŽ Structure data, graphs, social networks and multi-linked data
īŽ Heterogeneous databases and legacy databases
īŽ Spatial, spatiotemporal, multimedia, text and Web data
īŽ Software programs, scientific simulations
īŽ New and sophisticated applications
30. 30
Chapter 1. Introduction
īŽ Why Data Mining?
īŽ What Is Data Mining?
īŽ A Multi-Dimensional View of Data Mining
īŽ What Kind of Data Can Be Mined?
īŽ What Kinds of Patterns Can Be Mined?
īŽ What Technology Are Used?
īŽ What Kind of Applications Are Targeted?
īŽ Major Issues in Data Mining
īŽ A Brief History of Data Mining and Data Mining Society
īŽ Summary
31. 31
Applications of Data Mining
īŽ Web page analysis: from web page classification, clustering to
PageRank & HITS algorithms
īŽ Collaborative analysis & recommender systems
īŽ Basket data analysis to targeted marketing
īŽ Biological and medical data analysis: classification, cluster analysis
(microarray data analysis), biological sequence analysis, biological
network analysis
īŽ Data mining and software engineering (e.g., IEEE Computer, Aug.
2009 issue)
īŽ From major dedicated data mining systems/tools (e.g., SAS, MS SQL-
Server Analysis Manager, Oracle Data Mining Tools) to invisible data
mining
32. 32
Chapter 1. Introduction
īŽ Why Data Mining?
īŽ What Is Data Mining?
īŽ A Multi-Dimensional View of Data Mining
īŽ What Kind of Data Can Be Mined?
īŽ What Kinds of Patterns Can Be Mined?
īŽ What Technology Are Used?
īŽ What Kind of Applications Are Targeted?
īŽ Major Issues in Data Mining
īŽ A Brief History of Data Mining and Data Mining Society
īŽ Summary
33. 33
Major Issues in Data Mining (1)
īŽ Mining Methodology
īŽ Mining various and new kinds of knowledge
īŽ Mining knowledge in multi-dimensional space
īŽ Data mining: An interdisciplinary effort
īŽ Boosting the power of discovery in a networked environment
īŽ Handling noise, uncertainty, and incompleteness of data
īŽ Pattern evaluation and pattern- or constraint-guided mining
īŽ User Interaction
īŽ Interactive mining
īŽ Incorporation of background knowledge
īŽ Presentation and visualization of data mining results
34. 34
Major Issues in Data Mining (2)
īŽ Efficiency and Scalability
īŽ Efficiency and scalability of data mining algorithms
īŽ Parallel, distributed, stream, and incremental mining methods
īŽ Diversity of data types
īŽ Handling complex types of data
īŽ Mining dynamic, networked, and global data repositories
īŽ Data mining and society
īŽ Social impacts of data mining
īŽ Privacy-preserving data mining
īŽ Invisible data mining
35. 35
Chapter 1. Introduction
īŽ Why Data Mining?
īŽ What Is Data Mining?
īŽ A Multi-Dimensional View of Data Mining
īŽ What Kind of Data Can Be Mined?
īŽ What Kinds of Patterns Can Be Mined?
īŽ What Technology Are Used?
īŽ What Kind of Applications Are Targeted?
īŽ Major Issues in Data Mining
īŽ A Brief History of Data Mining and Data Mining Society
īŽ Summary
36. 36
A Brief History of Data Mining Society
īŽ 1989 IJCAI Workshop on Knowledge Discovery in Databases
īŽ Knowledge Discovery in Databases (G. Piatetsky-Shapiro and W. Frawley,
1991)
īŽ 1991-1994 Workshops on Knowledge Discovery in Databases
īŽ Advances in Knowledge Discovery and Data Mining (U. Fayyad, G.
Piatetsky-Shapiro, P. Smyth, and R. Uthurusamy, 1996)
īŽ 1995-1998 International Conferences on Knowledge Discovery in Databases
and Data Mining (KDDâ95-98)
īŽ Journal of Data Mining and Knowledge Discovery (1997)
īŽ ACM SIGKDD conferences since 1998 and SIGKDD Explorations
īŽ More conferences on data mining
īŽ PAKDD (1997), PKDD (1997), SIAM-Data Mining (2001), (IEEE) ICDM
(2001), etc.
īŽ ACM Transactions on KDD starting in 2007
37. 37
Conferences and Journals on Data Mining
īŽ KDD Conferences
īŽ ACM SIGKDD Int. Conf. on
Knowledge Discovery in
Databases and Data Mining (KDD)
īŽ SIAM Data Mining Conf. (SDM)
īŽ (IEEE) Int. Conf. on Data Mining
(ICDM)
īŽ European Conf. on Machine
Learning and Principles and
practices of Knowledge Discovery
and Data Mining (ECML-PKDD)
īŽ Pacific-Asia Conf. on Knowledge
Discovery and Data Mining
(PAKDD)
īŽ Int. Conf. on Web Search and
Data Mining (WSDM)
īŽ Other related conferences
īŽ DB conferences: ACM SIGMOD,
VLDB, ICDE, EDBT, ICDT, âĻ
īŽ Web and IR conferences: WWW,
SIGIR, WSDM
īŽ ML conferences: ICML, NIPS
īŽ PR conferences: CVPR,
īŽ Journals
īŽ Data Mining and Knowledge
Discovery (DAMI or DMKD)
īŽ IEEE Trans. On Knowledge and
Data Eng. (TKDE)
īŽ KDD Explorations
īŽ ACM Trans. on KDD
38. 38
Where to Find References? DBLP, CiteSeer, Google
īŽ Data mining and KDD (SIGKDD: CDROM)
īŽ Conferences: ACM-SIGKDD, IEEE-ICDM, SIAM-DM, PKDD, PAKDD, etc.
īŽ Journal: Data Mining and Knowledge Discovery, KDD Explorations, ACM TKDD
īŽ Database systems (SIGMOD: ACM SIGMOD AnthologyâCD ROM)
īŽ Conferences: ACM-SIGMOD, ACM-PODS, VLDB, IEEE-ICDE, EDBT, ICDT, DASFAA
īŽ Journals: IEEE-TKDE, ACM-TODS/TOIS, JIIS, J. ACM, VLDB J., Info. Sys., etc.
īŽ AI & Machine Learning
īŽ Conferences: Machine learning (ML), AAAI, IJCAI, COLT (Learning Theory), CVPR, NIPS, etc.
īŽ Journals: Machine Learning, Artificial Intelligence, Knowledge and Information Systems,
IEEE-PAMI, etc.
īŽ Web and IR
īŽ Conferences: SIGIR, WWW, CIKM, etc.
īŽ Journals: WWW: Internet and Web Information Systems,
īŽ Statistics
īŽ Conferences: Joint Stat. Meeting, etc.
īŽ Journals: Annals of statistics, etc.
īŽ Visualization
īŽ Conference proceedings: CHI, ACM-SIGGraph, etc.
īŽ Journals: IEEE Trans. visualization and computer graphics, etc.
39. 39
Chapter 1. Introduction
īŽ Why Data Mining?
īŽ What Is Data Mining?
īŽ A Multi-Dimensional View of Data Mining
īŽ What Kind of Data Can Be Mined?
īŽ What Kinds of Patterns Can Be Mined?
īŽ What Technology Are Used?
īŽ What Kind of Applications Are Targeted?
īŽ Major Issues in Data Mining
īŽ A Brief History of Data Mining and Data Mining Society
īŽ Summary
40. 40
Summary
īŽ Data mining: Discovering interesting patterns and knowledge from
massive amount of data
īŽ A natural evolution of database technology, in great demand, with
wide applications
īŽ A KDD process includes data cleaning, data integration, data
selection, transformation, data mining, pattern evaluation, and
knowledge presentation
īŽ Mining can be performed in a variety of data
īŽ Data mining functionalities: characterization, discrimination,
association, classification, clustering, outlier and trend analysis, etc.
īŽ Data mining technologies and applications
īŽ Major issues in data mining
41. 41
Recommended Reference Books
īŽ S. Chakrabarti. Mining the Web: Statistical Analysis of Hypertex and Semi-Structured Data. Morgan
Kaufmann, 2002
īŽ R. O. Duda, P. E. Hart, and D. G. Stork, Pattern Classification, 2ed., Wiley-Interscience, 2000
īŽ T. Dasu and T. Johnson. Exploratory Data Mining and Data Cleaning. John Wiley & Sons, 2003
īŽ U. M. Fayyad, G. Piatetsky-Shapiro, P. Smyth, and R. Uthurusamy. Advances in Knowledge Discovery and
Data Mining. AAAI/MIT Press, 1996
īŽ U. Fayyad, G. Grinstein, and A. Wierse, Information Visualization in Data Mining and Knowledge
Discovery, Morgan Kaufmann, 2001
īŽ J. Han and M. Kamber. Data Mining: Concepts and Techniques. Morgan Kaufmann, 3rd ed., 2011
īŽ D. J. Hand, H. Mannila, and P. Smyth, Principles of Data Mining, MIT Press, 2001
īŽ T. Hastie, R. Tibshirani, and J. Friedman, The Elements of Statistical Learning: Data Mining, Inference,
and Prediction, 2nd ed., Springer-Verlag, 2009
īŽ B. Liu, Web Data Mining, Springer 2006.
īŽ T. M. Mitchell, Machine Learning, McGraw Hill, 1997
īŽ G. Piatetsky-Shapiro and W. J. Frawley. Knowledge Discovery in Databases. AAAI/MIT Press, 1991
īŽ P.-N. Tan, M. Steinbach and V. Kumar, Introduction to Data Mining, Wiley, 2005
īŽ S. M. Weiss and N. Indurkhya, Predictive Data Mining, Morgan Kaufmann, 1998
īŽ I. H. Witten and E. Frank, Data Mining: Practical Machine Learning Tools and Techniques with Java
Implementations, Morgan Kaufmann, 2nd ed. 2005