Lecture 01 – Introduction to DM

ANAUM HAMID
 Gentle Reminder

“Switch Off” your Mobile Phone Or Switch

Mobile Phone to “Silent Mode”
 Contents
1 Why Data Mining?

2 A Multi-Dimensional View of Data Mining

3 What Kind of Data Can Be Mined?

4 What Kinds of Patterns Can Be Mined?

5 What Technology Are Used?

6 What Kind of Applications Are Targeted?

7 Major Issues in Data Mining

8 A Brief History of Data Mining and Data Mining Society

 Motivation
v  Data is Easily Collected…

4
 Motivation
v  …and Accessed

5
 Why Data Mining?

v  The Explosive Growth of Data: from terabytes Transaction

Databases
to petabytes Logs

§  Data collection and data availability

•  Automated data collection tools, Data
database systems, Web, computerized Streams

society
§  Major sources of abundant data Time series

•  Business: Web, e-commerce,

transactions, stocks, …
•  Science: Remote sensing, Social
Data Networks
bioinformatics, scientific simulation, …
warehouses
•  Society and everyone: news, digital
cameras, YouTube
6
 Data Tsunami is Coming

v Web and Social Networks kilobyte (kB) 103

megabyte (MB) 106
generates amount of data gigabyte (GB) 109
§  Google processes 100 PB per day, 3 terabyte (TB) 1012
million servers petabyte (PB) 1015
exabyte (EB) 1018
§  Facebook has 300 PB of user data
zettabyte (ZB) 1021
per day yottabyte (YB) 1024
§  YouTube has 1000PB video storage brontobyte(BB) 1027
§  235 TBs data collected by the US geopbyte(GPB) 1030
Library of Congress
§  15 out of 17 sectors in the US have
more data stored per company than
the US Library of Congress
7
2017 This is what happen in an Internet
Minute?

8
 Why Data Mining?

v We are drowning in

data, but starving for
knowledge!
v “Necessity is the
mother of invention”—
Data mining—
Automated analysis of
massive data sets

9
 Evolution of Database Technology
v  1960s:
§  Data collection, database creation, IMS and network DBMS
v  1970s:
§  Relational data model, relational DBMS implementation
v  1980s:
§  RDBMS, advanced data models (extended-relational, OO, deductive, etc.)
§  Application-oriented DBMS (spatial, scientific, engineering, etc.)
v  1990s:
§  Data mining, data warehousing, multimedia databases, and Web databases
v  2000s
§  Stream data management and mining
§  Data mining and its applications
§  Web technology (XML, data integration) and global information systems
v  2010s
§  Big Data
§  Cloud Computing
10
 Evolution of Database Technology
Data Collection & Database Creation
(1960s and earlier)
• Primitive file processing

Database Management Systems

(1970--‐early 1980s)
• Relational DB
• Query languages (SQL)
• Online Transaction Processing (OLTP)
•

Advanced Database Advanced Data Analysis Web-based Big Data & Cloud
Systems (late 1980s --- today) Databases (2010s and today)
(mid 1980s --‐ today) • Data warehousing (1990s--‐ • Big Data
• Advanced database & OLAP today) • Cloud computing
applications • Data mining • XML--‐based
•… & KD DB systems
•… •…
11
 Origins of Data Mining

v Draws ideas from machine learning/AI, pattern

recognition, statistics, and database systems
v Traditional techniques may be unsuitable due to data
that is
§  Large-scale
§  High dimensional
§  Heterogeneous
§  Complex
§  Distributed
v  A key component of the emerging field of data science and
data-driven discovery
Data Science
 What Is Data Mining?

v Data mining (knowledge discovery from data)

§  Extraction of interesting (non-trivial, implicit, previously
unknown and potentially useful) patterns or knowledge
from huge amount of data
§  Data mining: a misnomer?
v Alternative names
§  Knowledge discovery (mining) in databases (KDD),
knowledge extraction, data/pattern analysis, data
archeology, data dredging, information harvesting,
business intelligence, etc.
v Watch out: Is everything “data mining”?
§  Simple search and query processing
§  (Deductive) expert systems 14
 Knowledge Discovery (KDD) Process
v  This is a view from typical
database systems and data
Pattern Evaluation
warehousing communities
v  Data mining plays an essential
role in the knowledge discovery
process Data Mining

Task-relevant Data

Data Warehouse Selection

Data Cleaning

Data Integration

Databases 15
Example: A Web Mining Framework

16
Example: A Web Mining Framework
 Data Mining in Business Intelligence

Increasing potential
to support
business decisions End User
Decision
Making

Data Presentation Business

Analyst
Visualization Techniques
Data Mining Data
Information Discovery Analyst

Data Exploration
Statistical Summary, Querying, and Reporting

Data Preprocessing/Integration, Data Warehouses

DBA
Data Sources
Paper, Files, Web documents, Scientific experiments, Database Systems
18
Example: Mining vs. Data Exploration

v Business intelligence view

§  Warehouse, data cube, reporting but not much mining
v Business objects vs. data mining tools
v Supply chain example: tools
v Data presentation
v Exploration

19
 KDD Process: A Typical View from ML and
Statistics

Input Data Data Pre- Data Post-

Processing Mining Processing

Data integration Pattern discovery Pattern evaluation

Normalization Association & correlation Pattern selection
Feature selection Classification Pattern interpretation
Clustering
Dimension reduction Pattern visualization
Outlier analysis
…………

v  This is a view from typical machine learning and statistics communities

20
 Example: Medical Data Mining

v Health care & medical data mining – often

adopted such a view in statistics and machine
learning
v Preprocessing of the data (including feature
extraction and dimension reduction)
v Classification or/and clustering processes
v Post-processing for presentation

21
 Multi-Dimensional View of Data Mining
v  Data to be mined
§  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
v  Knowledge to be mined (or: Data mining functions)
§  Characterization, discrimination, association, classification,
clustering, trend/deviation, outlier analysis, etc.
§  Descriptive vs. predictive data mining
§  Multiple/integrated functions and mining at multiple levels
v  Techniques utilized
§  Data-intensive, data warehouse (OLAP), machine learning, statistics,
pattern recognition, visualization, high-performance, etc.
v  Applications adapted
§  Retail, telecommunication, banking, fraud analysis, bio-data mining,
stock market analysis, text mining, Web mining, etc. 22
Multi-Dimensional View of Data Mining
Data to be
mined

Knowledge
mined

Techniques
Applications

23
Multi-Dimensional View of Data Mining
Data to be mined:
Data to be
•  Database data;
•  Warehouse data; mined
•  Transactional data;
•  Stream;
Knowledge
•  Spatiotemporal; mined
•  Time‐series;
•  Sequence;
•  Text and web;
•  Multimedia;
•  Graphs;
•  Social networks;
•  …

Techniques
Applications

24
Multi-Dimensional View of Data Mining
Data to be
mined

Knowledge
mined

Knowledge mined:
•  Characterization;
•  Discrimination;
•  Association;
•  Classification;
•  Clustering;
•  Trend/deviation;
•  Outlier analysis;
•  Descriptive mining;
Techniques
•  Predictive mining;
Applications •  …

25
Multi-Dimensional View of Data Mining
Data to be
mined

Knowledge
mined
Techniques:
•  Data‐intensive;
•  OLAP;
•  Machine learning;
•  Statistics;
•  Pattern recognition;
•  Visualization;
•  High‐performance;
•  …
Techniques
Applications

26
 Multi-Dimensional View of Data Mining
Data to be
mined

Applications: Knowledge
•  Retail; mined
•  Telecommunication;
•  Banking;
•  Fraud analysis;
•  Bio‐data mining;
•  Stock market analysis;
•  Text mining;
•  Web mining;
•  …

Techniques
Applications

27
 Data Mining: On What Kinds of Data?
v  Database-oriented data sets and applications
§  Relational database, data warehouse,
transactional database
v  Advanced data sets and advanced applications
§  Data streams and sensor data
§  Time-series data, temporal data, sequence data
(incl. bio-sequences)
§  Structure data, graphs, social networks and multi-
linked data
§  Object-relational databases
§  Heterogeneous databases and legacy databases
§  Spatial data and spatiotemporal data
§  Multimedia database
§  Text databases
§  The World-Wide Web

28
 Data Mining Function: (1) Generalization

v Information integration and data warehouse construction

§  Data cleaning, transformation, integration, and
multidimensional data model
v Data cube technology
§  Scalable methods for computing (i.e., materializing)
multidimensional aggregates
§  OLAP (online analytical processing)
v Multidimensional concept description: Characterization
and discrimination
§  Generalize, summarize, and contrast data
characteristics, e.g., dry vs. wet region
29
Data Mining Function: (2) Association and
Correlation Analysis
v  Frequent patterns (or frequent itemsets)
§  What items are frequently purchased together
in your Walmart?
v  Association, correlation vs. causality
§  A typical association rule
•  Bread à Eggs [0.5%, 75%] (support,
confidence)
§  Are strongly associated items also strongly
correlated?
v  How to mine such patterns and rules efficiently in
large datasets?
v  How to use such patterns for classification,
clustering, and other applications?

30
 Data Mining Function: (3) Classification

v  Classification and label prediction

§  Construct models (functions) based on some training
examples
§  Describe and distinguish classes or concepts for future
prediction
•  E.g., classify countries based on (climate), or classify
cars based on (gas mileage)
§  Predict some unknown class labels
v  Typical methods
§  Decision trees, naïve Bayesian classification, support
vector machines, neural networks, rule-based
classification, pattern-based classification, logistic
regression, …
v  Typical applications:
§  Credit card fraud detection, direct marketing, classifying
stars, diseases, web-pages, … 31
Data Mining Function: (4) Cluster Analysis

v Unsupervised learning (i.e., Class label is unknown)

v Group data to form new categories (i.e., clusters), e.g.,
cluster houses to find distribution patterns
v Principle: Maximizing intra-class similarity & minimizing
interclass similarity
v Many methods and applications

32
Data Mining Function: (5) Outlier Analysis

v  Outlier analysis
§  Outlier: A data object that does not comply with the general
behavior of the data
§  Noise or exception? ― One person’s garbage could be another
person’s treasure
§  Methods: by product of clustering or regression analysis, …
§  Useful in fraud detection, rare events analysis

33
Data Mining Tasks: Pictorial Summary

34
Time and Ordering: Sequential Pattern, Trend
and Evolution Analysis
v Sequence, trend and evolution analysis
§  Trend, time-series, and deviation analysis: e.g.,
regression and value prediction
§  Sequential pattern mining
•  e.g., first buy digital camera, then buy large SD memory
cards
§  Periodicity analysis
§  Motifs and biological sequence analysis
•  Approximate and consecutive motifs
§  Similarity-based analysis
v Mining data streams
§  Ordered, time-varying, potentially infinite, data streams
35
 Structure and Network Analysis

v  Graph mining
§  Finding frequent subgraphs (e.g., chemical compounds), trees
(XML), substructures (web fragments)
v  Information network analysis
§  Social networks: actors (objects, nodes) and relationships (edges)
•  e.g., author networks in CS, terrorist networks
§  Multiple heterogeneous networks
•  A person could be multiple information networks: friends,
family, classmates, …
§  Links carry a lot of semantic information: Link mining
v  Web mining
§  Web is a big information network: from PageRank to Google
§  Analysis of Web information networks
•  Web community discovery, opinion mining, usage mining, …
36
 Evaluation of Knowledge
v Are all mined knowledge interesting?
§  One can mine tremendous amount of “patterns” and knowledge
§  Some may fit only certain dimension space (time, location, …)
§  Some may not be representative, may be transient, …
v Evaluation of mined knowledge → directly mine only
interesting knowledge?
§  Descriptive vs. predictive
§  Coverage
§  Typicality vs. novelty
§  Accuracy
§  Timeliness
§  … 37
 Data Mining: Confluence of Multiple Disciplines

Machine Pattern
Statistics
Learning Recognition

Applications Data Mining Visualization

Algorithm Database High-Performance

Technology Computing

38
 Why Confluence of Multiple Disciplines?

v Tremendous amount of data

§  Algorithms must be highly scalable to handle such as tera-bytes of
data
v High-dimensionality of data
§  Micro-array may have tens of thousands of dimensions
v High complexity of data
§  Data streams and sensor data
§  Time-series data, temporal data, sequence data
§  Structure data, graphs, social networks and multi-linked data
§  Heterogeneous databases and legacy databases
§  Spatial, spatiotemporal, multimedia, text and Web data
§  Software programs, scientific simulations
v New and sophisticated applications
39
 Applications of Data Mining
v  Web page analysis: from web page classification, clustering to
PageRank & HITS algorithms
v  Collaborative analysis & recommender systems
v  Basket data analysis to targeted marketing
v  Biological and medical data analysis: classification, cluster analysis
(microarray data analysis), biological sequence analysis, biological
network analysis
v  Data mining and software engineering (e.g., IEEE Computer, Aug.
2009 issue)
v  From major dedicated data mining systems/tools (e.g., SAS, MS SQL-
Server Analysis Manager, Oracle Data Mining Tools) to invisible data
mining

40
 Major Issues in Data Mining (1)

v  Mining Methodology
§  Mining various and new kinds of knowledge
§  Mining knowledge in multi-dimensional space
§  Data mining: An interdisciplinary effort
§  Boosting the power of discovery in a networked environment
§  Handling noise, uncertainty, and incompleteness of data
§  Pattern evaluation and pattern- or constraint-guided mining
v  User Interaction
§  Interactive mining
§  Incorporation of background knowledge
§  Presentation and visualization of data mining results

41
 Major Issues in Data Mining (2)

v  Efficiency and Scalability

§  Efficiency and scalability of data mining algorithms
§  Parallel, distributed, stream, and incremental mining methods
v  Diversity of data types
§  Handling complex types of data
§  Mining dynamic, networked, and global data repositories
v  Data mining and society
§  Social impacts of data mining
§  Privacy-preserving data mining
§  Invisible data mining

42
 A Brief History of Data Mining Society

v  1989 IJCAI Workshop on Knowledge Discovery in Databases

§  Knowledge Discovery in Databases (G. Piatetsky-Shapiro and W.
Frawley, 1991)
v  1991-1994 Workshops on Knowledge Discovery in Databases
§  Advances in Knowledge Discovery and Data Mining (U. Fayyad, G.
Piatetsky-Shapiro, P. Smyth, and R. Uthurusamy, 1996)
v  1995-1998 International Conferences on Knowledge Discovery in Databases
and Data Mining (KDD’95-98)
§  Journal of Data Mining and Knowledge Discovery (1997)
v  ACM SIGKDD conferences since 1998 and SIGKDD Explorations
v  More conferences on data mining
§  PAKDD (1997), PKDD (1997), SIAM-Data Mining (2001), (IEEE) ICDM
(2001), etc.
v  ACM Transactions on KDD starting in 2007
43
 Conferences and Journals on Data Mining

v  KDD Conferences n  Other related conferences

§  ACM SIGKDD Int. Conf. on
n  DB conferences: ACM SIGMOD,
Knowledge Discovery in
VLDB, ICDE, EDBT, ICDT, …
Databases and Data Mining
(KDD) n  Web and IR conferences: WWW,
§  SIAM Data Mining Conf. (SDM) SIGIR, WSDM
§  (IEEE) Int. Conf. on Data Mining n  ML conferences: ICML, NIPS
(ICDM) n  PR conferences: CVPR,
§  European Conf. on Machine n  Journals
Learning and Principles and
n  Data Mining and Knowledge
practices of Knowledge Discovery
and Data Mining (ECML-PKDD) Discovery (DAMI or DMKD)
§  Pacific-Asia Conf. on Knowledge n  IEEE Trans. On Knowledge and
Discovery and Data Mining Data Eng. (TKDE)
(PAKDD) n  KDD Explorations
§  Int. Conf. on Web Search and n  ACM Trans. on KDD
Data Mining (WSDM)
44
Where to Find References? DBLP, CiteSeer, Google

v  Data mining and KDD (SIGKDD: CDROM)

§  Conferences: ACM-SIGKDD, IEEE-ICDM, SIAM-DM, PKDD, PAKDD, etc.
§  Journal: Data Mining and Knowledge Discovery, KDD Explorations, ACM TKDD
v  Database systems (SIGMOD: ACM SIGMOD Anthology—CD ROM)
§  Conferences: ACM-SIGMOD, ACM-PODS, VLDB, IEEE-ICDE, EDBT, ICDT, DASFAA
§  Journals: IEEE-TKDE, ACM-TODS/TOIS, JIIS, J. ACM, VLDB J., Info. Sys., etc.
v  AI & Machine Learning
§  Conferences: Machine learning (ML), AAAI, IJCAI, COLT (Learning Theory), CVPR, NIPS,
etc.
§  Journals: Machine Learning, Artificial Intelligence, Knowledge and Information Systems,
IEEE-PAMI, etc.
v  Web and IR
§  Conferences: SIGIR, WWW, CIKM, etc.
§  Journals: WWW: Internet and Web Information Systems,
v  Statistics
§  Conferences: Joint Stat. Meeting, etc.
§  Journals: Annals of statistics, etc.
v  Visualization
§  Conference proceedings: CHI, ACM-SIGGraph, etc.
§  Journals: IEEE Trans. visualization and computer graphics, etc. 45
 Summary

v  Data mining: Discovering interesting patterns and knowledge from

massive amount of data
v  A natural evolution of database technology, in great demand, with
wide applications
v  A KDD process includes data cleaning, data integration, data
selection, transformation, data mining, pattern evaluation, and
knowledge presentation
v  Mining can be performed in a variety of data
v  Data mining functionalities: characterization, discrimination,
association, classification, clustering, outlier and trend analysis, etc.
v  Data mining technologies and applications
v  Major issues in data mining

46
 Recommended Reference Books
v  S. Chakrabarti. Mining the Web: Statistical Analysis of Hypertex and Semi-Structured Data. Morgan Kaufmann,
2002
v  R. O. Duda, P. E. Hart, and D. G. Stork, Pattern Classification, 2ed., Wiley-Interscience, 2000
v  T. Dasu and T. Johnson. Exploratory Data Mining and Data Cleaning. John Wiley & Sons, 2003
v  U. M. Fayyad, G. Piatetsky-Shapiro, P. Smyth, and R. Uthurusamy. Advances in Knowledge Discovery and Data
Mining. AAAI/MIT Press, 1996
v  U. Fayyad, G. Grinstein, and A. Wierse, Information Visualization in Data Mining and Knowledge Discovery,
Morgan Kaufmann, 2001
v  J. Han and M. Kamber. Data Mining: Concepts and Techniques. Morgan Kaufmann, 3rd ed., 2011
v  D. J. Hand, H. Mannila, and P. Smyth, Principles of Data Mining, MIT Press, 2001
v  T. Hastie, R. Tibshirani, and J. Friedman, The Elements of Statistical Learning: Data Mining, Inference, and
Prediction, 2nd ed., Springer-Verlag, 2009
v  B. Liu, Web Data Mining, Springer 2006.
v  T. M. Mitchell, Machine Learning, McGraw Hill, 1997
v  G. Piatetsky-Shapiro and W. J. Frawley. Knowledge Discovery in Databases. AAAI/MIT Press, 1991
v  P.-N. Tan, M. Steinbach and V. Kumar, Introduction to Data Mining, Wiley, 2005
v  S. M. Weiss and N. Indurkhya, Predictive Data Mining, Morgan Kaufmann, 1998
v  I. H. Witten and E. Frank, Data Mining: Practical Machine Learning Tools and Techniques with Java
Implementations, Morgan Kaufmann, 2nd ed. 2005

47
 TUTORIAL 01

1.  Give a brief definition for the term “Data Mining”?

2.  What is data mining? In your answer, address the following:
a)  Is it another hype?
b)  Is it a simple transformation of technology developed from databases, statistics,
and machine learning?
c)  Explain how the evolution of database technology led to data mining.
d)  Describe the steps involved in data mining when viewed as a process of
knowledge discovery.
3.  What does this statement mean “Drowning in data starving for
knowledge”?
4.  Why data mining is called misnomer?
5.  List down alternative names of data mining.
6.  Explain the difference between “Explorative Data Mining” and
“Predictive Data Mining” and give one example of each.

49
 TUTORIAL 01

6.  State three different applications for which data mining techniques
seem appropriate. Informally explain each application.
7.  Suppose your task as a software engineer at Big-University is to
design a data mining system to examine their university course
database, which contains the following information: the name,
address, and status (e.g., undergraduate or graduate) of each
student, the courses taken, and their cumulative grade point average
(GPA). Describe the architecture you would choose. What is the
purpose of each component of this architecture?
8.  What are the major challenges of mining a huge amount of data
(such as billions of tuples) in comparison with mining a small amount
of data (such as a few hundred tuple data set)?

50