Contents

Preface

Acknowledgments

Authors

Introduction

1  Introduction to Machine Learning

1.1  Introduction

1.2  Preliminaries

1.2.1  Machine Learning: Where Several Disciplines Meet

1.2.2  Supervised Learning

1.2.3  Unsupervised Learning

1.2.4  Semi-Supervised Learning

1.2.5  Reinforcement Learning

1.2.6  Validation and Evaluation

1.3  Applications of Machine Learning Algorithms

1.3.1  Automatic Recognition of Handwritten Postal Codes

1.3.2  Computer-Aided Diagnosis

1.3.3  Computer Vision

1.3.3.1  Driverless Cars

1.3.3.2  Face Recognition and Security

1.3.4  Speech Recognition

1.3.5  Text Mining

1.3.5.1  Where Text and Image Data Can Be Used Together

1.4  The Present and the Future

1.4.1  Thinking Machines

1.4.2  Smart Machines

1.4.3  Deep Blue

1.4.4  IBM’s Watson

1.4.5  Google Now

1.4.6  Apple’s Siri

1.4.7  Microsoft’s Cortana

1.5  Objective of This Book

References

SECTION I SUPERVISED LEARNING ALGORITHMS

2  Decision Trees

2.1  Introduction

2.2  Entropy

2.2.1  Example

2.2.2  Understanding the Concept of Number of Bits

2.3  Attribute Selection Measure

2.3.1  Information Gain of ID3

2.3.2  The Problem with Information Gain

2.4  Implementation in MATLAB^®

2.4.1  Gain Ratio of C4.5

2.4.2  Implementation in MATLAB

References

3  Rule-Based Classifiers

3.1  Introduction to Rule-Based Classifiers

3.2  Sequential Covering Algorithm

3.3  Algorithm

3.4  Visualization

3.5  Ripper

3.5.1  Algorithm

3.5.2  Understanding Rule Growing Process

3.5.3  Information Gain

3.5.4  Pruning

3.5.5  Optimization

References

4  Naïve Bayesian Classification

4.1  Introduction

4.2  Example

4.3  Prior Probability

4.4  Likelihood

4.5  Laplace Estimator

4.6  Posterior Probability

4.7  MATLAB Implementation

References

5  The k-Nearest Neighbors Classifiers

5.1  Introduction

5.2  Example

5.3  k-Nearest Neighbors in MATLAB^®

References

6  Neural Networks

6.1  Perceptron Neural Network

6.1.1  Perceptrons

6.2  MATLAB Implementation of the Perceptron Training and Testing Algorithms

6.3  Multilayer Perceptron Networks

6.4  The Backpropagation Algorithm

6.4.1  Weights Updates in Neural Networks

6.5  Neural Networks in MATLAB

References

7  Linear Discriminant Analysis

7.1  Introduction

7.2  Example

References

8  Support Vector Machine

8.1  Introduction

8.2  Definition of the Problem

8.2.1  Design of the SVM

8.2.2  The Case of Nonlinear Kernel

8.3  The SVM in MATLAB^®

References

SECTION II UNSUPERVISED LEARNING ALGORITHMS

9  k-Means Clustering

9.1  Introduction

9.2  Description of the Method

9.3  The k-Means Clustering Algorithm

9.4  The k-Means Clustering in MATLAB^®

10  Gaussian Mixture Model

10.1  Introduction

10.2  Learning the Concept by Example

References

11  Hidden Markov Model

11.1  Introduction

11.2  Example

11.3  MATLAB Code

References

12  Principal Component Analysis

12.1  Introduction

12.2  Description of the Problem

12.3  The Idea behind the PCA

12.3.1  The SVD and Dimensionality Reduction

12.4  PCA Implementation

12.4.1  Number of Principal Components to Choose

12.4.2  Data Reconstruction Error

12.5  The Following MATLAB^® Code Applies the PCA

12.6  Principal Component Methods in Weka

12.7  Example: Polymorphic Worms Detection Using PCA

12.7.1  Introduction

12.7.2  SEA, MKMP, and PCA

12.7.3  Overview and Motivation for Using String Matching

12.7.4  The KMP Algorithm

12.7.5  Proposed SEA

12.7.6  An MKMP Algorithm

12.7.6.1  Testing the Quality of the Generated Signature for Polymorphic Worm A

12.7.7  A Modified Principal Component Analysis

12.7.7.1  Our Contributions in the PCA

12.7.7.2  Testing the Quality of Generated Signature for Polymorphic Worm A

12.7.7.3  Clustering Method for Different Types of Polymorphic Worms

12.7.8  Signature Generation Algorithms Pseudo-Codes

12.7.8.1  Signature Generation Process

References

Appendix I:   Transcript of Conversations with Chatbot

Appendix II: Creative Chatbot

Index