Bayesian Networks, the result of the convergence of artificial intelligence with statistics, are growing in popularity. Their versatility and modelling power is now employed across a variety of fields for the purposes of analysis, simulation, prediction and diagnosis. This book provides a general introduction to Bayesian networks, defining and illustrating the basic concepts with pedagogical examples and twenty real-life case studies drawn from a range of fields including medicine, computing, natural sciences and engineering. Designed to help analysts, engineers, scientists and professionals taking part in complex decision processes to successfully implement Bayesian networks, this book equips readers with proven methods to generate, calibrate, evaluate and validate Bayesian networks. The book: Provides the tools to overcome common practical challenges such as the treatment of missing input data, interaction with experts and decision makers, determination of the optimal granularity and size of the model.  Highlights the strengths of Bayesian networks whilst also presenting a discussion of their limitations. Compares Bayesian networks with other modelling techniques such as neural networks, fuzzy logic and fault trees. Describes, for ease of comparison, the main features of the major Bayesian network software packages: Netica, Hugin, Elvira and Discoverer, from the point of view of the user. Offers a historical perspective on the subject and analyses future directions for research. Written by leading experts with practical experience of applying Bayesian networks in finance, banking, medicine, robotics, civil engineering, geology, geography, genetics, forensic science, ecology, and industry, the book has much to offer both practitioners and researchers involved in statistical analysis or modelling in any of these fields.

Table of Contents:
Foreword ix Preface xi 1 Introduction to Bayesian networks 1 1.1 Models 1 1.2 Probabilistic vs. deterministic models 5 1.3 Unconditional and conditional independence 9 1.4 Bayesian networks 11 2 Medical diagnosis 15 2.1 Bayesian networks in medicine 15 2.2 Context and history 17 2.3 Model construction 19 2.4 Inference 26 2.5 Model validation 28 2.6 Model use 30 2.7 Comparison to other approaches 31 2.8 Conclusions and perspectives 32 3 Clinical decision support 33 3.1 Introduction 33 3.2 Models and methodology 34 3.3 The Busselton network 35 3.4 The PROCAM network 40 3.5 The PROCAM Busselton network 44 3.6 Evaluation 46 3.7 The clinical support tool: TakeHeartII 47 3.8 Conclusion 51 4 Complex genetic models 53 4.1 Introduction 53 4.2 Historical perspectives 54 4.3 Complex traits 56 4.4 Bayesian networks to dissect complex traits 59 4.5 Applications 64 4.6 Future challenges 71 5 Crime risk factors analysis 73 5.1 Introduction 73 5.2 Analysis of the factors affecting crime risk 74 5.3 Expert probabilities elicitation 75 5.4 Data preprocessing 76 5.5 A Bayesian network model 78 5.6 Results 80 5.7 Accuracy assessment 83 5.8 Conclusions 84 6 Spatial dynamics in France 87 6.1 Introduction 87 6.2 An indicator-based analysis 89 6.3 The Bayesian network model 97 6.4 Conclusions 109 7 Inference problems in forensic science 113 7.1 Introduction 113 7.2 Building Bayesian networks for inference 116 7.3 Applications of Bayesian networks in forensic science 120 7.4 Conclusions 126 8 Conservation of marbled murrelets in British Columbia 127 8.1 Context/history 127 8.2 Model construction 129 8.3 Model calibration, validation and use 136 8.4 Conclusions/perspectives 147 9 Classifiers for modeling of mineral potential 149 9.1 Mineral potential mapping 149 9.2 Classifiers for mineral potential mapping 151 9.3 Bayesian network mapping of base metal deposit 157 9.4 Discussion 166 9.5 Conclusions 171 10 Student modeling 173 10.1 Introduction 173 10.2 Probabilistic relational models 175 10.3 Probabilistic relational student model 176 10.4 Case study 180 10.5 Experimental evaluation 182 10.6 Conclusions and future directions 185 11 Sensor validation 187 11.1 Introduction 187 11.2 The problem of sensor validation 188 11.3 Sensor validation algorithm 191 11.4 Gas turbines 197 11.5 Models learned and experimentation 198 11.6 Discussion and conclusion 202 12 An information retrieval system 203 12.1 Introduction 203 12.2 Overview 205 12.3 Bayesian networks and information retrieval 206 12.4 Theoretical foundations 207 12.5 Building the information retrieval system 215 12.6 Conclusion 223 13 Reliability analysis of systems 225 13.1 Introduction 225 13.2 Dynamic fault trees 227 13.3 Dynamic Bayesian networks 228 13.4 A case study: The Hypothetical Sprinkler System 230 13.5 Conclusions 237 14 Terrorism risk management 239 14.1 Introduction 240 14.2 The Risk Influence Network 250 14.3 Software implementation 254 14.4 Site Profiler deployment 259 14.5 Conclusion 261 15 Credit-rating of companies 263 15.1 Introduction 263 15.2 Naive Bayesian classifiers 264 15.3 Example of actual credit-ratings systems 264 15.4 Credit-rating data of Japanese companies 266 15.5 Numerical experiments 267 15.6 Performance comparison of classifiers 273 15.7 Conclusion 276 16 Classification of Chilean wines 279 16.1 Introduction 279 16.2 Experimental setup 281 16.3 Feature extraction methods 285 16.4 Classification results 288 16.5 Conclusions 298 17 Pavement and bridge management 301 17.1 Introduction 301 17.2 Pavement management decisions 302 17.3 Bridge management 307 17.4 Bridge approach embankment – case study 308 17.5 Conclusion 312 18 Complex industrial process operation 313 18.1 Introduction 313 18.2 A methodology for Root Cause Analysis 314 18.3 Pulp and paper application 321 18.4 The ABB Industrial IT platform 325 18.5 Conclusion 326 19 Probability of default for large corporates 329 19.1 Introduction 329 19.2 Model construction 332 19.3 BayesCredit 335 19.4 Model benchmarking 341 19.5 Benefits from technology and software 342 19.6 Conclusion 343 20 Risk management in robotics 345 20.1 Introduction 345 20.2 DeepC 346 20.3 The ADVOCATE II architecture 352 20.4 Model development 354 20.5 Model usage and examples 360 20.6 Benefits from using probabilistic graphical models 361 20.7 Conclusion 362 21 Enhancing Human Cognition 365 21.1 Introduction 365 21.2 Human foreknowledge in everyday settings 366 21.3 Machine foreknowledge 369 21.4 Current application and future research needs 373 21.5 Conclusion 375 22 Conclusion 377 22.1 An artificial intelligence perspective 377 22.2 A rational approach of knowledge 379 22.3 Future challenges 384 Bibliography 385 Index 427