Bayesian Reasoning and Gaussian Processes for Machine Learning Applications

Description

This book introduces Bayesian reasoning and Gaussian processes into machine learning applications. Bayesian methods are applied in many areas, such as game development, decision making, and drug discovery. It is very effective for machine learning algorithms in handling missing data and extracting information from small datasets.Bayesian Reasoning and Gaussian Processes for Machine Learning Applications uses a statistical background to understand continuous distributions and how learning can be viewed from a probabilistic framework. The chapters progress into such machine learning topics as belief network and Bayesian reinforcement learning, which is followed by Gaussian process introduction, classification, regression, covariance, and performance analysis of Gaussian processes with other models. Cover Half Title Title Page Copyright Page Table of Contents Preface Editors Contributors 1. Introduction to Naive Bayes and a Review on Its Subtypes with Applications 1.1 Introduction 1.2 Intuition behind the Naive Bayes Algorithm and Its Subtypes with Applications 1.2.1 Why Is It Called Naive Bayes? 1.2.2 Bayes Theorem – Intuition behind the Classification 1.2.2.1 Bayes Theorem 1.2.2.2 Bayes Theorem in Machine Learning 1.2.3 Types of Naive Bayes Models 1.2.4 Gaussian Naive Bayes 1.2.5 Predictions Using Gaussian Naive Bayes Model 1.2.6 Bernoulli Classification 1.2.6.1 Bernoulli Statistics or Distribution 1.2.6.2 Rule for Bernoulli Naive Bayes Classifier 1.2.6.3 An Example for Bernoulli Naive Bayes 1.2.6.4 Advantages 1.2.6.5 Disadvantages 1.2.7 Multinomial Naive Bayes Classifier 1.2.8 Differences between Gaussian, Bernoulli, and Multinomial Distributions 1.2.9 Advantages of Naive Bayes 1.2.10 Disadvantages of Naive Bayes 1.3 Real-Time Application: Human Activity Recognition Using Naive Bayes Algorithm 1.3.1 Dataset Attributes 1.3.2 Naive Bayes Algorithm–Based Result 1.4 Conclusion References 2. A Review on the Different Regression Analysis in Supervised Learning 2.1 Introduction 2.2 Linear Regression 2.2.1 Simple Linear Regression 2.2.2 Finding the Line Equation for the Data 2.2.3 Multivariable Linear Regression 2.3 Logistic Regression 2.3.1 Logistic Function (Sigmoid Function) 2.3.2 Logistic Regression Equation 2.3.3 Types of Logistic Regressions 2.4 Regularization 2.4.1 Ridge Regression (L2) 2.4.2 Lasso Regression (L1) 2.4.3 Lasso Regression’s Drawbacks 2.5 Polynomial Regression 2.6 Bayesian Regression References 3. Methods to Predict the Performance Analysis of Various Machine Learning Algorithms 3.1 Introduction 3.2 Analysis of Algorithms 3.3 Evaluation of Performance in Machine Learning Models 3.3.1 Methods for Model Evaluation 3.3.1.1 Confusion Matrix 3.3.1.2 Accuracy 3.3.1.3 Precision 3.3.1.4 Recall 3.3.1.5 Specificity 3.3.1.6 F-score 3.3.1.7 ROC Curve 3.4 Evaluation of Performance of Regression Model 3.4.1 R Square or Adjusted R Square 3.4.2 Mean Square Error or Root Mean Square Error 3.4.3 Mean Absolute Error 3.5 Examples 3.5.1 Coding Example of Evaluation of Performance in Machine Learning Models 3.5.2 Coding Example of Evaluation of Performance of Regression Model References 4. A Viewpoint on Belief Networks and Their Applications 4.1 Introduction 4.2 Belief Networks Designing 4.3 Applications of Belief Networks References 5. Reinforcement Learning Using Bayesian Algorithms with Applications 5.1 Introduction 5.1.1 Applications of Reinforcement Learning 5.2 Bayesian Reinforcement Learning 5.3 Model-Free Reinforcement Learning 5.4 Value-Function-Based Algorithms References 6. Alerting System for Gas Leakage in Pipelines 6.1 Introduction 6.2 Previous Work 6.2.1 Machine Learning and Acoustic Method Applied to Leak Detection and Location in Low-Pressure Gas Pipelines 6.2.2 Detection and Online Prediction of Leak Magnitude in a Gas Pipeline Using an Acoustic Method and Neural Network Data Processing 6.2.3 Experimental Study on Leak Detection and Location for Gas Pipeline Based on Acoustic Method 6.2.4 Detection of Leak Acoustic Signal in Buried Gas Pipe Based on the Time–Frequency Analysis 6.2.5 Leakage Detection and Prediction of Location in a Smart Water Grid Using SVM Classification 6.2.6 Leakage Detection of a Spherical Water Storage Tank in a Chemical Industry Using Acoustic Emissions 6.3 Methodology/Proposed Work 6.3.1 Machine Learning in Leak Detection 6.3.2 Importance of Sensors and Sensors Used 6.3.2.1 Pressure Sensor 6.3.2.2 Temperature Sensor 6.3.2.3 Proximity Sensors 6.3.3 How Does It Work? 6.4 Linear Regression Algorithm 6.5 Scikit-Learn Library 6.6 Conclusion References 7. Two New Nonparametric Models for Biological Networks 7.1 Introduction 7.2 Methods 7.2.1 Random Forest Algorithm 7.2.2 Gaussian Graphical Model 7.2.3 Multivariate Adaptive Regression Splines 7.3 Application 7.3.1 Application via Random Forest Algorithm 7.3.1.1 Application via Simulated Data 7.3.1.2 Application via Real Data 7.3.2 Application via MARS and Gaussian Graphical Model 7.4 Conclusion Acknowledgments References 8. Generating Various Types of Graphical Models via MARS 8.1 Introduction 8.2 Lasso-Based MARS and the Relation with GGM 8.3 Applications 8.4 Conclusion References 9. Financial Applications of Gaussian Processes and Bayesian Optimization 9.1 Introduction 9.2 Gaussian Processes 9.2.1 Gaussian Process Definition 9.2.2 Gaussian Process Regression 9.2.3 Covariance Functions 9.2.4 Hyperparameter Selection 9.2.5 Classification 9.3 Bayesian Optimization 9.3.1 General Principles 9.4 Financial Applications 9.4.1 Yield Curve Modeling 9.4.2 Portfolio Optimization 9.4.2.1 Trend Following Strategy 9.4.2.2 Hyperparameter Assessment of the Trend-Following Strategy 9.5 Summary References 10. Bayesian Network Inference on Diabetes Risk Prediction Data 10.1 Introduction 10.2 Methods 10.2.1 Score-Based BN Inference Algorithms 10.2.2 Constraint-Based Algorithms (Markov Blanket Learning Algorithms) 10.3 Results 10.4 Discussion References Index