MATLAB Machine Learning Recipes: A Problem-Solution Approach

Harness the power of MATLAB to resolve a wide range of machine learning challenges. This new and updated third edition provides examples of technologies critical to machine learning. Each example solves a real-world problem, and all code provided is executable. You can easily look up a particular problem and follow the steps in the solution. This book has something for everyone interested in machine learning. It also has material that will allow those with an interest in other technology areas to see how machine learning and MATLAB can help them solve problems in their areas of expertise. The chapter on data representation and MATLAB graphics includes new data types and additional graphics. Chapters on fuzzy logic, simple neural nets, and autonomous driving have new examples added. And there is a new chapter on spacecraft attitude determination using neural nets. Authors Michael Paluszek and Stephanie Thomas show how all of these technologies allow you to build sophisticated applications to solve problems with pattern recognition, autonomous driving, expert systems, and much more. What You Will Learn Write code for machine learning, adaptive control, and estimation using MATLAB Use MATLAB graphics and visualization tools for machine learning Become familiar with neural nets Build expert systems Understand adaptive control Gain knowledge of Kalman Filters Who This Book Is For Software engineers, control engineers, university faculty, undergraduate and graduate students, hobbyists. Contents About the Authors About the Technical Reviewer Introduction 1 An Overview of Machine Learning 1.1 Introduction 1.2 Elements of Machine Learning 1.2.1 Data 1.2.2 Models 1.2.3 Training Supervised Learning Unsupervised Learning Semi-supervised Learning Online Learning 1.3 The Learning Machine 1.4 Taxonomy of Machine Learning 1.5 Control 1.5.1 Kalman Filters 1.5.2 Adaptive Control 1.6 Autonomous Learning Methods 1.6.1 Regression 1.6.2 Decision Trees 1.6.3 Neural Networks Introduction Generative Deep Learning Reinforcement Learning 1.6.4 Support Vector Machines (SVMs) 1.7 Artificial Intelligence 1.7.1 What Is Artificial Intelligence? 1.7.2 Intelligent Cars 1.7.3 Expert Systems 1.8 Summary 2 Data for Machine Learning in MATLAB 2.1 Introduction to MATLAB Data Types 2.1.1 Matrices 2.1.2 Cell Arrays 2.1.3 Data Structures 2.1.4 Numerics 2.1.5 Images 2.1.6 Datastore 2.1.7 Tall Arrays 2.1.8 Sparse Matrices 2.1.9 Tables and Categoricals 2.1.10 Large MAT-Files 2.2 Initializing a Data Structure 2.2.1 Problem 2.2.2 Solution 2.2.3 How It Works 2.3 mapreduce on an Image Datastore 2.3.1 Problem 2.3.2 Solution 2.3.3 How It Works 2.4 Processing Table Data 2.4.1 Problem 2.4.2 Solution 2.4.3 How It Works 2.5 String Concatenation 2.5.1 Problem 2.5.2 Solution 2.5.3 How It Works 2.6 Arrays of Strings 2.6.1 Problem 2.6.2 Solution 2.6.3 How It Works 2.7 Substrings 2.7.1 Problem 2.7.2 Solution 2.7.3 How It Works 2.8 Reading an Excel Spreadsheet into a Table 2.8.1 Problem 2.8.2 Solution 2.8.3 How It Works 2.9 Accessing ChatGPT 2.9.1 Problem 2.9.2 Solution 2.9.3 How It Works 2.10 Summary 3 MATLAB Graphics 3.1 2D Line Plots 3.1.1 Problem 3.1.2 Solution 3.1.3 How It Works 3.2 General 2D Graphics 3.2.1 Problem 3.2.2 Solution 3.2.3 How It Works 3.3 Custom Two-Dimensional Diagrams 3.3.1 Problem 3.3.2 Solution 3.3.3 How It Works 3.4 Three-Dimensional Box 3.4.1 Problem 3.4.2 Solution 3.4.3 How It Works 3.5 Draw a 3D Object with a Texture 3.5.1 Problem 3.5.2 Solution 3.5.3 How It Works 3.6 General 3D Graphics 3.6.1 Problem 3.6.2 Solution 3.6.3 How It Works 3.7 Building a GUI 3.7.1 Problem 3.7.2 Solution 3.7.3 How It Works 3.8 Animating a Bar Chart 3.8.1 Problem 3.8.2 Solution 3.8.3 How It Works 3.9 Drawing a Robot 3.9.1 Problem 3.9.2 Solution 3.9.3 How It Works 3.10 Importing a Model 3.10.1 Problem 3.10.2 Solution 3.10.3 How It Works 3.11 Summary 4 Kalman Filters 4.1 Gaussian Distribution 4.2 A State Estimator Using a Linear Kalman Filter 4.2.1 Problem 4.2.2 Solution 4.2.3 How It Works 4.3 Using the Extended Kalman Filter for State Estimation 4.3.1 Problem 4.3.2 Solution 4.3.3 How It Works 4.4 Using the UKF for State Estimation 4.4.1 Problem 4.4.2 Solution 4.4.3 How It Works 4.5 Using the UKF for Parameter Estimation 4.5.1 Problem 4.5.2 Solution 4.5.3 How It Works 4.6 Range to a Car 4.6.1 Problem 4.6.2 Solution 4.6.3 How It Works 4.7 Summary 5 Adaptive Control 5.1 Self-Tuning: Tuning an Oscillator 5.1.1 Problem 5.1.2 Solution 5.1.3 How It Works 5.2 Implement MRAC 5.2.1 Problem 5.2.2 Solution 5.2.3 How It Works 5.3 Generating a Square Wave Input 5.3.1 Problem 5.3.2 Solution 5.3.3 How It Works 5.4 Demonstrate MRAC for a Rotor 5.4.1 Problem 5.4.2 Solution 5.4.3 How It Works 5.5 Ship Steering: Implement Gain Scheduling for Steering Control of a Ship 5.5.1 Problem 5.5.2 Solution 5.5.3 How It Works 5.6 Spacecraft Pointing 5.6.1 Problem 5.6.2 Solution 5.6.3 How It Works 5.7 Direct Adaptive Control 5.7.1 Problem 5.7.2 Solution 5.7.3 How It Works 5.8 Summary 6 Fuzzy Logic 6.1 Building Fuzzy Logic Systems 6.1.1 Problem 6.1.2 Solution 6.1.3 How It Works 6.2 Implement Fuzzy Logic 6.2.1 Problem 6.2.2 Solution 6.2.3 How It Works 6.3 Window Wiper Fuzzy Controller 6.3.1 Problem 6.3.2 Solution 6.3.3 How It Works 6.4 Simple Discrete HVAC Fuzzy Controller 6.4.1 Problem 6.4.2 Solution 6.4.3 How It Works 6.5 Variable HVAC Fuzzy Controller 6.5.1 Problem 6.5.2 Solution 6.5.3 How It Works 6.6 Summary 7 Neural Aircraft Control 7.1 Longitudinal Motion 7.1.1 Problem 7.1.2 Solution 7.1.3 How It Works 7.2 Numerically Finding Equilibrium 7.2.1 Problem 7.2.2 Solution 7.2.3 How It Works 7.3 Numerical Simulation of the Aircraft 7.3.1 Problem 7.3.2 Solution 7.3.3 How It Works 7.4 Activation Function 7.4.1 Problem 7.4.2 Solution 7.4.3 How It Works 7.5 Neural Net for Learning Control 7.5.1 Problem 7.5.2 Solution 7.5.3 How It Works 7.6 Enumeration of All Sets of Inputs 7.6.1 Problem 7.6.2 Solution 7.6.3 How It Works 7.7 Write a Sigma-Pi Neural Net Function 7.7.1 Problem 7.7.2 Solution 7.7.3 How It Works 7.8 Implement PID Control 7.8.1 Problem 7.8.2 Solution 7.8.3 How It Works 7.9 PID Control of Pitch 7.9.1 Problem 7.9.2 Solution 7.9.3 How It Works 7.10 Neural Net for Pitch Dynamics 7.10.1 Problem 7.10.2 Solution 7.10.3 How It Works 7.11 Nonlinear Simulation 7.11.1 Problem 7.11.2 Solution 7.11.3 How It Works 7.12 Summary 8 Introduction to Neural Nets 8.1 Daylight Detector 8.1.1 Problem 8.1.2 Solution 8.1.3 How It Works 8.2 Modeling a Pendulum 8.2.1 Problem 8.2.2 Solution 8.2.3 How It Works 8.3 Single Neuron Angle Estimator 8.3.1 Problem 8.3.2 Solution 8.3.3 How It Works 8.4 Designing a Neural Net for the Pendulum 8.4.1 Problem 8.4.2 Solution 8.4.3 How It Works 8.5 XOR Example 8.6 Training 8.7 Summary 9 Classification of Numbers Using Neural Networks 9.1 Generate Test Images with Defects 9.1.1 Problem 9.1.2 Solution 9.1.3 How It Works 9.2 Create the Neural Net Functions 9.2.1 Problem 9.2.2 Solution 9.2.3 How It Works 9.3 Train a Network with One Output Node 9.3.1 Problem 9.3.2 Solution 9.3.3 How It Works 9.4 Testing the Neural Network 9.4.1 Problem 9.4.2 Solution 9.4.3 How It Works 9.5 Train a Network with Many Outputs 9.5.1 Problem 9.5.2 Solution 9.5.3 How It Works 9.6 Summary 10 Data Classification with Decision Trees 10.1 Generate Test Data 10.1.1 Problem 10.1.2 Solution 10.1.3 How It Works 10.2 Drawing Trees 10.2.1 Problem 10.2.2 Solution 10.2.3 How It Works 10.3 Implementation 10.3.1 Problem 10.3.2 Solution 10.3.3 How It Works 10.4 Creating a Tree 10.4.1 Problem 10.4.2 Solution 10.4.3 How It Works 10.5 Handmade Tree 10.5.1 Problem 10.5.2 Solution 10.5.3 How It Works 10.6 Training and Testing 10.6.1 Problem 10.6.2 Solution 10.6.3 How It Works 10.7 Summary 11 Pattern Recognition with Deep Learning 11.1 Obtain Data Online for Training a Neural Net 11.1.1 Problem 11.1.2 Solution 11.1.3 How It Works 11.2 Generating Training Images of Cats 11.2.1 Problem 11.2.2 Solution 11.2.3 How It Works 11.3 Matrix Convolution 11.3.1 Problem 11.3.2 Solution 11.3.3 How It Works 11.4 Convolution Layer 11.4.1 Problem 11.4.2 Solution 11.4.3 How It Works 11.5 Pooling to Outputs of a Layer 11.5.1 Problem 11.5.2 Solution 11.5.3 How It Works 11.6 Fully Connected Layer 11.6.1 Problem 11.6.2 Solution 11.6.3 How It Works 11.7 Determining the Probability 11.7.1 Problem 11.7.2 Solution 11.7.3 How It Works 11.8 Test the Neural Network 11.8.1 Problem 11.8.2 Solution 11.8.3 How It Works 11.9 Recognizing an Image 11.9.1 Problem 11.9.2 Solution 11.9.3 How It Works 11.10 Using AlexNet 11.10.1 Problem 11.10.2 Solution 11.10.3 How It Works Summary 12 Multiple Hypothesis Testing 12.1 Overview 12.2 Theory 12.2.1 Introduction 12.2.2 Example 12.2.3 Algorithm 12.2.4 Measurement Assignment and Tracks 12.2.5 Hypothesis Formation 12.2.6 Track Pruning 12.3 Billiard Ball Kalman Filter 12.3.1 Problem 12.3.2 Solution 12.3.3 How It Works 12.4 Billiard Ball MHT 12.4.1 Problem 12.4.2 Solution 12.4.3 How It Works 12.5 One-Dimensional Motion 12.5.1 Problem 12.5.2 Solution 12.5.3 How It Works 12.6 One-Dimensional MHT 12.6.1 Problem 12.6.2 Solution 12.6.3 How It Works 12.7 Summary 13 Autonomous Driving with MHT 13.1 Automobile Dynamics 13.1.1 Problem 13.1.2 Solution 13.1.3 How It Works 13.2 Automobile Radar 13.2.1 Problem 13.2.2 Solution 13.2.3 How It Works 13.3 Passing Control 13.3.1 Problem 13.3.2 Solution 13.3.3 How It Works 13.4 Automobile Animation 13.4.1 Problem 13.4.2 Solution 13.4.3 How It Works 13.4.4 Solution 13.5 Automobile Simulation and the Kalman Filter 13.5.1 Problem 13.5.2 Solution 13.5.3 How It Works 13.6 Automobile Target Tracking 13.6.1 Problem 13.6.2 Solution 13.6.3 How It Works 13.7 Summary 14 Spacecraft Attitude Determination 14.1 Star Catalog 14.1.1 Problem 14.1.2 Solution 14.1.3 How It Works 14.2 Camera Model 14.2.1 Problem 14.2.2 Solution 14.2.3 How It Works 14.3 Celestial Sphere 14.3.1 Problem 14.3.2 Solution 14.3.3 How It Works 14.4 Attitude Simulation of Camera Views 14.4.1 Problem 14.4.2 Solution 14.4.3 How It Works 14.5 Yaw Angle Rotation 14.5.1 Problem 14.5.2 Solution 14.5.3 How It Works 14.6 Yaw Images 14.6.1 Problem 14.6.2 Solution 14.6.3 How It Works 14.7 Attitude Determination 14.7.1 Problem 14.7.2 Solution 14.7.3 How It Works 14.8 Summary 15 Case-Based Expert Systems 15.1 Building Expert Systems 15.1.1 Problem 15.1.2 Solution 15.1.3 How It Works 15.2 Running an Expert System 15.2.1 Problem 15.2.2 Solution 15.2.3 How It Works 15.3 Summary A A Brief History A.1 Introduction A.2 Artificial Intelligence A.3 Learning Control A.4 Machine Learning A.5 Generative Machine Learning A.6 Reinforcement Learning A.7 The Future B Software for Machine Learning B.1 Autonomous Learning Software B.2 Commercial MATLAB Software B.2.1 MathWorks Products Statistics and Machine Learning Toolbox Optimization Toolbox Global Optimization Toolbox Text Analytics Toolbox Deep Learning Toolbox B.2.2 Princeton Satellite Systems Products Core Control Toolbox Target Tracking B.3 Non-MATLAB Products for Machine Learning B.3.1 R B.3.2 scikit-learn B.3.3 LIBSVM B.4 Products for Optimization B.4.1 LOQO B.4.2 SNOPT B.4.3 GLPK B.4.4 CVX B.4.5 SeDuMi B.4.6 YALMIP B.5 Products for Expert Systems B.6 MATLAB mex Files B.6.1 Problem B.6.2 Solution B.6.3 How It Works Bibliography Index