Neural networks for conditional probability estimation : forecasting beyond point predictions

Chapter 1 Introduction

Chapter 2 A Universal Approximator Network for Predicting Conditional Probability Densities

Chapter 3 A Maximum Likelihood Training Scheme

Chapter 4 Benchmark Problems

Chapter 5 Demonstration of the Model Performance on the Benchmark Problems

Chapter 6 Random Vector Functional Link (RVFL) Networks

Chapter 7 Improved Training Scheme Combining the Expectation Maximisation (EM) Algorithm with the RVFL Approach

Chapter 8 Empirical Demonstration: Combining EM and RVFL

Chapter 9 A simple Bayesian regularisation scheme

Chapter 11 The Bayesian Evidence Scheme for Model Selection

Chapter 12 Demonstration of the Bayesian Evidence Scheme for Regularisation

Chapter 13 Network Committees and Weighting Schemes

Chapter 14 Demonstration: Committees of Networks Trained with Different Regularisation Schemes

Chapter 15 Automatic Relevance Determination (ARD)

Chapter 16 A Real-World Application: The Boston Housing Data

Chapter 17 Summary

Chapter 18 Appendix: Derivation of the Hessian for the Bayesian Evidence Scheme