Dunne A Statistical Approach to Neural Networks for Pattern Recognition

1. Auflage 2008
ISBN: 978-0-470-14814-3
Verlag: John Wiley & Sons
Format: PDF
Kopierschutz: Adobe DRM (»Systemvoraussetzungen)

E-Book, Englisch, 288 Seiten, E-Book

Reihe: Wiley Series in Computational Statistics

An accessible and up-to-date treatment featuring the connectionbetween neural networks and statistics
A Statistical Approach to Neural Networks for PatternRecognition presents a statistical treatment of the MultilayerPerceptron (MLP), which is the most widely used of the neuralnetwork models. This book aims to answer questions that arise whenstatisticians are first confronted with this type of model, suchas:
How robust is the model to outliers?
Could the model be made more robust?
Which points will have a high leverage?
What are good starting values for the fitting algorithm?
Thorough answers to these questions and many more are included,as well as worked examples and selected problems for the reader.Discussions on the use of MLP models with spatial and spectral dataare also included. Further treatment of highly important principalaspects of the MLP are provided, such as the robustness of themodel in the event of outlying or atypical data; the influence andsensitivity curves of the MLP; why the MLP is a fairly robustmodel; and modifications to make the MLP more robust. The authoralso provides clarification of several misconceptions that areprevalent in existing neural network literature.
Throughout the book, the MLP model is extended in severaldirections to show that a statistical modeling approach can makevaluable contributions, and further exploration for fitting MLPmodels is made possible via the R and S-PLUS® codes that areavailable on the book's related Web site. A Statistical Approach toNeural Networks for Pattern Recognition successfully connectslogistic regression and linear discriminant analysis, thus makingit a critical reference and self-study guide for students andprofessionals alike in the fields of mathematics, statistics,computer science, and electrical engineering.

Autoren/Hrsg.

Dunne, Robert A.

Inhaltsverzeichnis

Notation and Code Examples.
Preface.
Acknowledgments.
1. Introduction.
2. The Multi-Layer Perception Model.
3. Linear Discriminant Analysis.
4. Activation and Penalty Functions.
5. Model Fitting and Evaluation.
6. The Task-Based MLP.
7. Incorporating Spatial Information into an MLP Classifier.
8. Influence Curves for the Multi-Layer PerceptronClassifier.
9. The Sensitivity Curves of the MLP Classifier.
10. A Robust Fitting Procedure for MLP Models.
11. Smoothed Weights.
12. Translation Invariance.
13. Fixed-slope Training.
Appendix A. Function Minimization.
Appendix B. Maximum Values of the Influence Curve.
Topic Index.

Über Autor(innen)

Robert A. Dunne, PhD, is Research Scientist in the Mathematical and Information Sciences Division of the Commonwealth Scientific and Industrial Research Organization (CSIRO) in North Ryde, Australia. Dr. Dunne received his PhD from Murdoch University, and his research interests include remote sensing and bioinformatics.