ISBN-10:
0130950831
ISBN-13:
9780130950833
Pub. Date:
12/04/1997
Publisher:
Pearson Education
Solving Data Mining Problems Through Pattern Recognition / Edition 1

Solving Data Mining Problems Through Pattern Recognition / Edition 1

Multimedia Set

Current price is , Original price is $73.32. You

Temporarily Out of Stock Online

Please check back later for updated availability.

This item is available online through Marketplace sellers.

Overview

Apply pattern recognition to find the hidden gems in your data!


Data mining technology is helping businesses everywhere to work smarter by revealing unknown patterns within existing archives. Applying the latest advances in pattern recognition software can give you a key competitive edge across all data mining applications. The tutorials and software package included in Solving Data Mining Problems through Pattern Recognition take advantage of machine learning techniques and neural networks to help you get the most out of your data. Besides explaining the most current theories, Solving Data Mining Problems through Pattern Recognition takes a practical approach to overall project development concerns.


The rigorous, multi-step method includes:



  • Defining the pattern recognition problem

  • Collection, preparation, and preprocessing of data

  • Choosing the appropriate algorithm and tuning algorithm parameters

Training, testing, and troubleshooting.


Pattern classification, estimation, and modeling are addressed using the following algorithms:



  • Linear and logistic regression

  • Unimodal Gaussian and Gaussian mixture

  • Multilayered perceptron/backpropagation and radial basis function neural networks

  • K nearest neighbors and nearest cluster

  • K means clustering.


While some aspects of pattern recognition involve advanced mathematical principles, most successful projects rely on a strong element of human experience and intuition. Solving Data Mining Problems through Pattern Recognition provides a strong theoretical grounding for beginners, yet it also contains detailed models and insights into real-world problem-solving that will inspire more experienced users, be they database designers, modelers, or project leaders.


This book includes a free, 90-day trial copy of Pattern Recognition Workbench, a powerful, easy-to-use system that combines machine learning, neural networks, and statistical algorithms to help you apply pattern recognition to your data right now. The enclosed CD-ROM runs under Windows(r) 95 and Windows NT(tm).


Product Details

ISBN-13: 9780130950833
Publisher: Pearson Education
Publication date: 12/04/1997
Series: Data Warehousing Institute Series from Prentice Hall PTR Series
Edition description: BK&CD ROM
Pages: 400
Product dimensions: 7.26(w) x 9.54(h) x 1.35(d)

Read an Excerpt

PREFACE: Preface

Data Mining

Data mining is a term usually applied to techniques that can be used to find underlying structure and relationships in large amounts of data. These techniques are drawn primarily from the related fields of neural networks, statistics, pattern classification, and machine learning. They are becoming more important as computer automation spreads and as the processing and storage capabilities of computers increase. Widely available, low-cost computer technology now makes it possible to both collect historical data and also to institute on-line analysis and controls for newly arriving data.

Applications

Data mining techniques are being successfully used for many diverse applications. These include paper and sheet metal production control, medical diagnosis and risk prediction, credit-card fraud detection, computer security break-in and misuse detection, computer user identity verification, aluminum and steel smelting control, oil refinery control, pollution control in power plants, fraudulent income tax return detection, automobile engine control and fault detection, electric motor fault and failure prediction, mass mailing and telemarketing, and simplifying world-wide-web usage by predicting useful sites from past user behavior. Benefits of Data Mining

Benefits in these and other applications include reduced costs due to more accurate control, more accurate future predictions, more effective fault detection and prediction, fraud detection and control, and automation of repetitive human tasks. In addition, services can be improved and extended due to a better understanding of underlyingprocesses and human behavior. Outline of this Book

This book provides a concise introduction to some of the most important input-output mapping, prediction, pattern classification, and clustering algorithms useful for data mining. This introduction is based on many collective years of experience by the authors, which has led to a focus on practical issues that must be addressed to successfully solve data mining problems. The book provides a basic road map for experts who know much about a specific application, but little about neural networks, statistics, pattern classification, or machine learning.

This road map first helps potential users determine whether input-output mapping, prediction, pattern classification, or clustering algorithms are appropriate for a given application. It then helps users determine which measurements, attributes, or features might be useful as inputs to these algorithms and provides guidance in collecting and formatting this data for computer analysis. Guidelines are then presented for accurately accessing performance using separate training, evaluation, and test data partitions or cross-validation. Finally, each important algorithm is described and guidance is provided concerning settings for parameters used to control the many algorithms. Multi-Algorithm Approach

An important truism presented in this book is that data mining is an art and that there is no single simple approach that is best for all problems. Rather, there are many algorithms and data representations, and the best strategy is to interactively experiment to find an approach that works for a particular data set. This human interaction is greatly simplified by the availability of software toolkits which allow users to interactively explore many algorithms on a common data set using the same performance metrics. This book focuses on one comprehensive software toolkit (Pattern Recognition Workbench) that includes most of the algorithms described and has the capability of handling large data sets. Details concerning this software, however, are relegated to the Appendix and to sections at the ends of chapters. These details can thus be skipped or used as examples of the types of information required to apply the various algorithms. Intended Audience

This book is most useful for persons who have a specific application in mind, but who know little about data mining algorithms. They can use this book to determine whether the algorithms presented can be applied to their application, to learn terminology, and to provide guidance when they try out some of the recommended approaches using a software toolkit. More experienced users who want to understand the theory behind prediction, mapping, control, pattern classification, and clustering or who would like to read detailed descriptions concerning specific data mining applications should explore other more advanced texts.

Richard P. Lippmann

Table of Contents

List of Figures
xv(4)
List of Tables
xix(2)
Forward xxi(2)
Preface xxiii
Chapter 1 Introduction
1.1 Pattern Recognition by Humans
1-2(1)
1.2 Pattern Recognition by Computers
1-3(1)
1.3 Data Mining and Pattern Recognition
1-3(2)
1.4 Types of Pattern Recognition
1-5(1)
1.5 Classification
1-6(2)
1.5.1 Calculation in Classification
1-6(1)
1.5.2 Uncertainty in Classification
1-6(1)
1.5.3 Computer-Automated Classification
1-7(1)
1.6 Estimation
1-8(2)
1.6.1 Calculation in Estimation
1-8(1)
1.6.2 Uncertainty in Estimation
1-9(1)
1.6.3 Computer-Automated Estimation
1-10(1)
1.7 Developing a Model
1-10(10)
1.7.1 Fixed Models
1-10(2)
1.7.2 Parametric Models
1-12(3)
1.7.3 Nonparametric Models
1-15(1)
1.7.4 Preprocessing
1-16(2)
1.7.5 A Continuum of Methods
1-18(1)
1.7.6 Biases Due to Prior Knowledge
1-19(1)
1.8 The Purpose of this Book
1-20
Chapter 2 Key Concepts: Estimation
2.1 Terminology and Notation
2-2(2)
2.2 Characteristics of an Optimal Model
2-4(1)
2.3 Sources of Error
2-5(1)
2.4 Fixed Models
2-6(1)
2.5 Parametric Models
2-7(7)
2.5.1 Example: Linear Regression
2-8(1)
2.5.2 Generalization
2-9(3)
2.5.3 Shortcomings of Parametric Methods
2-12(1)
2.5.4 Iteration through Parametric Forms
2-13(1)
2.6 Nonparametric Models
2-14(13)
2.6.1 The Underlying Modeling Problem
2-14(1)
2.6.2 Heuristics in Nonparametric Modeling
2-15(3)
2.6.3 Approximation Architectures
2-18(3)
2.6.4 A Practical Nonparametric Approach
2-21(4)
2.6.5 The Role of Preprocessing
2-25(2)
2.7 Statistical Considerations
2-27
Chapter 3 Key Concepts: Classification
3.1 Terminology and Notation
3-2(1)
3.2 Characteristics of an Optimal Classifier
3-3(1)
3.3 Types of Models
3-4(5)
3.3.1 Decision-Region Boundaries
3-5(1)
3.3.2 Probability Density Functions
3-5(1)
3.3.3 Posterior Probabilities
3-6(3)
3.4 Approaches to Modeling
3-9(6)
3.4.1 Fixed Models
3-9(1)
3.4.2 Parametric Models
3-10(1)
3.4.3 Nonparametric Models
3-11(1)
3.4.4 The Role of Preprocessing
3-12(1)
3.4.5 The Importance of Multiple Techniques
3-13(2)
3.5 Statistical Considerations
3-15
Chapter 4 Additional Application Areas
4.1 Database Marketing
4-2(5)
4.1.1 Response Modeling
4-2(2)
4.1.2 Cross Selling
4-4(3)
4.2 Time-Series Prediction
4-7(1)
4.3 Detection
4-8(1)
4.4 Probability Estimation
4-9(1)
4.5 Information Compression
4-10(1)
4.6 Sensitivity Analysis
4-11
Chapter 5 Overview of the Development Process
5.1 Defining the Pattern Recognition Problem
5-3(1)
5.2 Collecting Data
5-3(1)
5.3 Preparing Data
5-3(1)
5.4 Preprocessing
5-4(1)
5.5 Selecting an Algorithm and Training Parameters
5-4(1)
5.6 Training and Testing
5-5(1)
5.7 Iterating Steps and Trouble-Shooting
5-6(1)
5.8 Evaluating the Final Model
5-6
Chapter 6 Defining the Pattern Recognition Problem
6.1 What Problems Are Suitable for Data-Driven Solutions?
6-2(2)
6.2 How Do You Evaluate Results?
6-4(1)
6.3 Is It a Classification or Estimation Problem?
6-4(3)
6.4 What Are the Inputs and Outputs?
6-7(6)
Apdx. Defining the Problem in PRW
6-13
Chapter 7 Collecting Data
7.1 What Data to Collect
7-1(2)
7.2 How to Collect Data
7-3(3)
7.3 How Much Data Is Enough
7-6(2)
7.4 Using Simulated Data
7-8(3)
Apdx. Importing Data into PRW
7-11
Chapter 8 Preparing Data
8.1 Missing Data Values
8-2(3)
8.2 Transforming Data into Numerical Values
8-5(4)
8.3 Inconsistent Data and Outliers
8-9(4)
Apdx. Preparing Data in PRW
8-13
A.1 Handling Missing Data
8-13(4)
A.2 Converting Non-Numeric Inputs
8-17(3)
A.3 Handling Inconsistent Data or Outliers
8-20
Chapter 9 Data Preprocessing
9.1 Why Should You Preprocess Your Data?
9-2(2)
9.2 Averaging Data Values
9-4(1)
9.3 Thresholding Data
9-5(1)
9.4 Reducing the Input Space
9-6(5)
9.5 Normalizing Data
9-11(6)
9.5.1 Why Normalize Data?
9-11(1)
9.5.2 Types of Normalization
9-12(5)
9.6 Modifying Prior Probabilities
9-17(1)
9.7 Other Considerations
9-18(3)
Apdx. Preprocessing in PRW
9-21
A.1 Averaging Time-Series Data
9-21(2)
A.2 Thresholding and Replacing Input Values
9-23(1)
A.3 Reducing the Input Space
9-23(5)
A.4 Normalizing Data
9-28(1)
A.5 Modifying Prior Input Probabilities
9-29
Chapter 10 Selecting Architectures and Training Parameters
10.1 Types of Algorithms
10-2(5)
10.2 How to Pick an Algorithm
10-7(2)
10.3 Practical Constraints
10-9(4)
10.3.1 Memory Usage
10-10(1)
10.3.2 Training Times
10-11(1)
10.3.3 Classification/Estimation Times
10-12(1)
10.4 Algorithm Descriptions
10-13(49)
10.4.1 Linear Regression
10-13(1)
10.4.2 Logistic Regression
10-14(4)
10.4.3 Unimodal Gaussian
10-18(5)
10.4.4 Multilayered Perceptron/Backpropagation
10-23(9)
10.4.5 Radial Basis Functions
10-32(5)
10.4.6 K Nearest Neighbors
10-37(4)
10.4.7 Gaussian Mixture
10-41(6)
10.4.8 Nearest Cluster
10-47(3)
10.4.9 K Means Clustering
10-50(4)
10.4.10 Decision Trees
10-54(6)
10.4.11 Other Nonparametric Architectures
10-60(2)
10.5 Algorithm Comparison Summary
10-62(3)
Apdx. Selecting Algorithms and Training Parameters in PRW
10-65
A.1 Selecting an Algorithm in PRW
10-65(2)
A.2 Setting Algorithm Parameters
10-67(1)
A.3 Linear Regression
10-67(1)
A.4 Logistic Regression
10-68(1)
A.5 Unimodal Gaussian
10-69(1)
A.6 Backpropagation/MLP
10-70(1)
A.7 Radial Basis Functions
10-71(1)
A.8 K Nearest Neighbors
10-72(1)
A.9 Gaussian Mixture
10-73(1)
A.10 Nearest Cluster
10-74(1)
A.11 K Means Clustering
10-75
Chapter 11 Training and Testing
11.1 Train, Test, and Evaluation Sets
11-1(3)
11.2 Validation Techniques
11-4(9)
11.2.1 Cross Validation
11-5(3)
11.2.2 Bootstrap Validation
11-8(2)
11.2.3 Sliding Window Validation
11-10(3)
Apdx. Training, Testing, and Reporting in PRW
11-13
A.1 The Experiment Manager
11-13(1)
A.2 Running Experiments
11-14(1)
A.3 Enabling and Disabling Experiments
11-15(2)
A.4 Scheduling Experiments
11-17(1)
A.5 Selecting Report Options
11-17(7)
A.6 Viewing Different Reports
11-24(1)
A.7 Cross Validation
11-25(3)
A.8 Sliding Window Validation
11-28
Chapter 12 Iterating Steps and Trouble-Shooting
12.1 Iterating to Improve Your Solution
12-1(4)
12.2 Automated Searches
12-5(9)
12.2.1 Input Variable Selection
12-6(6)
12.2.2 Algorithm Parameter Searches
12-12(2)
12.3 Trouble-Shooting
12-14(13)
12.3.1 Training Error Is High
12-14(2)
12.3.2 Test Error Is High
12-16(2)
12.3.3 Classification Problem Performs Poorly on Some Classes
12-18(1)
12.3.4 Problems with Production Accuracy
12-19(1)
12.3.5 Decision Tree Works Best by Far
12-19(1)
12.3.6 Backpropagation Does Not Converge
12-20(1)
12.3.7 Backpropagation Finds a Local Minimum Solution.
12-21(1)
12.3.8 Matrix Inversion Problem
12-22(1)
12.3.9 Unimodal Gaussian Has High Training Error
12-23(1)
12.3.10 Gaussian Mixture Diverges
12-24(1)
12.3.11 RBF Has High Training Error
12-24(3)
Apdx. Iterating in PRW
12-27
A.1 Overview of PRW Features
12-27(2)
A.2 Creating Multiple Spreadsheets
12-29(1)
A.3 Creating Multiple Experiment Managers
12-30(1)
A.4 Using Multiple Work Sessions
12-30(3)
A.5 Using Automated Searches
12-33(7)
A.6 Preprocessing Data
12-40(5)
A.7 Exporting Experiments and Reports
12-45(1)
A.8 Re-Using Experiment Parameters
12-46(1)
A.9 Building User Functions
12-46
Appendix A References and Suggested Reading
Appendix B Pattern Recognition Workbench
Appendix C Unica Technologies, Inc.
C.1 About Unica C-1(1)
C.2 Unica's Software Products C-2
Appendix D Glossary
Index Index-1
Software License Agreement
What's on this CD

Preface

PREFACE: Preface

Data Mining

Data mining is a term usually applied to techniques that can be used to find underlying structure and relationships in large amounts of data. These techniques are drawn primarily from the related fields of neural networks, statistics, pattern classification, and machine learning. They are becoming more important as computer automation spreads and as the processing and storage capabilities of computers increase. Widely available, low-cost computer technology now makes it possible to both collect historical data and also to institute on-line analysis and controls for newly arriving data.

Applications

Data mining techniques are being successfully used for many diverse applications. These include paper and sheet metal production control, medical diagnosis and risk prediction, credit-card fraud detection, computer security break-in and misuse detection, computer user identity verification, aluminum and steel smelting control, oil refinery control, pollution control in power plants, fraudulent income tax return detection, automobile engine control and fault detection, electric motor fault and failure prediction, mass mailing and telemarketing, and simplifying world-wide-web usage by predicting useful sites from past user behavior. Benefits of Data Mining

Benefits in these and other applications include reduced costs due to more accurate control, more accurate future predictions, more effective fault detection and prediction, fraud detection and control, and automation of repetitive human tasks. In addition, services can be improved and extended due to a better understanding ofunderlyingprocesses and human behavior. Outline of this Book

This book provides a concise introduction to some of the most important input-output mapping, prediction, pattern classification, and clustering algorithms useful for data mining. This introduction is based on many collective years of experience by the authors, which has led to a focus on practical issues that must be addressed to successfully solve data mining problems. The book provides a basic road map for experts who know much about a specific application, but little about neural networks, statistics, pattern classification, or machine learning.

This road map first helps potential users determine whether input-output mapping, prediction, pattern classification, or clustering algorithms are appropriate for a given application. It then helps users determine which measurements, attributes, or features might be useful as inputs to these algorithms and provides guidance in collecting and formatting this data for computer analysis. Guidelines are then presented for accurately accessing performance using separate training, evaluation, and test data partitions or cross-validation. Finally, each important algorithm is described and guidance is provided concerning settings for parameters used to control the many algorithms. Multi-Algorithm Approach

An important truism presented in this book is that data mining is an art and that there is no single simple approach that is best for all problems. Rather, there are many algorithms and data representations, and the best strategy is to interactively experiment to find an approach that works for a particular data set. This human interaction is greatly simplified by the availability of software toolkits which allow users to interactively explore many algorithms on a common data set using the same performance metrics. This book focuses on one comprehensive software toolkit (Pattern Recognition Workbench) that includes most of the algorithms described and has the capability of handling large data sets. Details concerning this software, however, are relegated to the Appendix and to sections at the ends of chapters. These details can thus be skipped or used as examples of the types of information required to apply the various algorithms. Intended Audience

This book is most useful for persons who have a specific application in mind, but who know little about data mining algorithms. They can use this book to determine whether the algorithms presented can be applied to their application, to learn terminology, and to provide guidance when they try out some of the recommended approaches using a software toolkit. More experienced users who want to understand the theory behind prediction, mapping, control, pattern classification, and clustering or who would like to read detailed descriptions concerning specific data mining applications should explore other more advanced texts.

Richard P. Lippmann

Customer Reviews

Most Helpful Customer Reviews

See All Customer Reviews