Transcript Chapter 1 - Cengage Learning

Chapter One
Introduction
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Chapter Overview
•
•
•
•
•
•
•
•
Roles of data, information and knowledge
Background of data mining
What is data mining?
Main data mining objectives
Data mining and other related disciplines
Current state of data mining
Promises and challenges
A brief preview of data mining tool Weka
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Data, Information and Knowledge
• Data (D)
– Isolated factual recording of separate objects
and events
– Enables the recording of the seen events
• Information (I)
– Fact of meaningful context represented by
relationships between isolated data items
– Information enables the responding to the
seen events
• Knowledge (K)
– Verified known information that is
accommodated into the business process
– Enable the anticipation of the unseen events
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
K
I
D
Data Mining: The Background
• Computerisation of operations in commercial,
governmental and scientific organisations has
resulted in large volumes of operational data, e.g.
– Itemised telephone bills
– Bank statements
– Supermarket transactions
– Share prices
– Scientific experimental data sets
– Published web pages
– CCTV video footages
– ……
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Data Mining: The Background
• Facts:
– Storing the data is an operational necessity
– Storing the data has become easy and affordable
– Data acquisition is fully or partially automatic and fast
• Consequences:
– The speed of data comprehension does not match the
speed of data acquisition
– Many commercial database management systems
(DBMSs) are not equipped with data comprehension
and analysis tools.
– We may be data rich, but information poor.
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Data Mining: The Background
• An intriguing quotable quote:
“I know half the money I spend on
advertising is wasted, but I can
never find out which half!”
Lord Leverhulme
President of Unilever
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Data Mining: What it is
Knowledge discovery in databases (KDD) refers to the efficient process
of searching through large volumes of raw data in databases to find
potentially useful information that is implicitly embedded in the data. Data
Mining is an integral step of KDD that discovers hidden patterns from an
input data set.
• Useful information; leading to a course of action or an
understanding of data
• Non-trivial implicit information; not the raw data, nor
the result of a simple data summary
• Real life databases; not laboratory generated data sets
• Efficient novel discovery methods; expected to be
scaled up and applied to large databases
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Data Mining: Useful Information
Example 1 (A well-known example, not a joke):
Customers who purchase beer are also likely (say 90%) to
purchase nappies.
Example 2 (May already be in practical use in credit card
applications):
If 20,000  Customer’s Salary  40,000 pounds and
Customer has a house, then Customer is a safe
customer.
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Data Mining: Non-trivial Information
Data retrieval
Online analytic processing
• Retrieval of stored data • Interactive reporting
• Trivial data aggregation
on stored data
• Written in standard SQL • Summarisation and
drilling along different
attributes
• Written in extended
SQL
Data mining
• Discovery of hidden and
embedded patterns
• Discovery algorithms
• Written in programming
language probably with
the assistance of SQL
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
High end of
sophistication
Low end of
sophistication
• Putting the “search for information” into a
spectrum:
Data Mining: Real-life Databases
• Characteristics of a real-life database
– The size may be extremely large
– The dimensionality can be very high
– Attributes can be of different data types
– Data quality can be very poor
– Data may exist in pieces and isolated in different
systems
– Value distribution can be extremely skewed
– Database content can be dynamic and evolving
– Data may lack traditional record-based structure
– Data are available on second storage media
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Data Mining: Efficient Algorithms
• Discovering interesting patterns supported by given facts
can be computationally hard because many discoveries
are combinatorial problems. Trivial algorithms may take
too long.
• A discovery algorithm is considered efficient if its
execution time and memory requirement are comparable
to those of sorting algorithms; otherwise, it is unlikely to
scale up well enough to cope with data sets of large
sizes.
• Efficient discovery algorithms may be hard to find. Using
advanced hardware, optimising the implementation of the
algorithms and developing approximate solutions can be
viable alternative options.
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Data Mining Objectives
• Classification
– Using existing data to form a classification model and then
using the model to assign an appropriate class label for a
data record (e.g. safe vs. risky customers)
• Estimation
– Similar to classification but to assign a value to an output
variable of a data record (e.g. estimated house value)
• Prediction
– Similar to classification and estimation, but more concerned
with future outcome of the output (e.g. tomorrow’s weather)
• Description
– General description of data characteristics (e.g. customer
profile)
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Data Mining & Other Disciplines
Machine Learning
Statistics
(Artificial Intelligence)
Inductive & deductive
learning methods
Data analysis theories
methods and measures
DATA MINING
Fast storage structures &
retrieval operations
Database
Management
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Data Mining: Current State
• Many data mining algorithms have been developed or
adapted
• Many data mining software tools have been built and
are in use
• A cross-industry methodology has been formed
• Besides general solutions, more application-oriented
data mining solutions are being developed
• More and more organisations are either doing their
own data mining or hiring consultants to do the job
• Data mining has been extended to web mining and
text mining
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Data Mining: Current State
• Some nuisances
– Mining cookies
– Spyware and miningware
– Intrusion to privacy
• Some serious problems
– “Big Brother is watching”
– Unfair advantages in trading practice e.g. highfrequency trading (HFT)
– Abuse of personal data
– Ethical concerns
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Data Mining: Promises
• Areas of data mining application:
– Finance and insurance
– Marketing and sales
– Medicine
– Agriculture
– Society, politics and economics
– Science
– Engineering
– Law enforcement
– Military and intelligence (classified)
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Data Mining: Challenges Faced
• Some difficult problems to solve
– Extremely large data sets
– Extremely high dimensionalities (curse of dimensions)
– Combinatorial problems and fast algorithms
– Meaningful evaluation of the patterns
– Discovery of changing and evolving patterns
– Integration of data mining techniques
– Comprehensibility of patterns
– Data pre-processing
– Mining non-standard complex data such as multimedia
materials
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Weka: A Brief Introduction
• Overview
– Java tool set developed at Univ. of Waikato (NZ)
– Free to download and used by many
– A wide range of learning and data pre-processing
methods and algorithms, with Java API
– Offering a GUI (Explorer) and a command-line (Simple
CLI) interface to the tools
– Experimenter module to assist the evaluation of
classification techniques
– KnowledgeFlow module to enable batch-processing
style discovery and incremental mining
– Some visualisation facilities
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Weka: A Brief Introduction
• Weka Explorer
– For investigative interactive data mining with small size data
sets
– Preprocess, Classify, Cluster, Associate, Select Attributes
and Visualise pages
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Weka: A Brief Introduction
• Weka Simple CLI
– Weka facilities as Java classes
– Calling the Java functions as commands
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Weka: A Brief Introduction
• Weka Experimenter
– Comparing performances of different classification solutions
on a collection of data sets
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Weka: A Brief Introduction
• Weka KnowledgeFlow
– Setting up a flow of knowledge discovery in a diagram
– Overview of the entire discovery project
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
Chapter Summary
• Importance of data in operation and importance of
information and knowledge in decision-making
• Data rich does not mean information rich
• Data mining: automatic or semi automatic data
understanding and decision support
• To classify, to estimate, to predict and to describe
• Data mining closely relates to database, statistics and
machine learning
• Data mining: from technology towards application
• A lot of potential uses and a lot of challenges to face
• Weka: excellent tool to support teaching data mining
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning
References
Read Chapter 1 of Data Mining Techniques and
Applications
Useful further references
• Han & Kamber, Chapter 1
• Berry & Linoff, Chapter 1 (business-like)
• Kdnuggets: http://www.kdnuggets.com/
Data Mining Techniques and Applications, 1st edition
Hongbo Du
ISBN 978-1-84480-891-5 © 2010 Cengage Learning