Overview
- Provides unique, comprehensive coverage of generalisation and regularisation: Provides the first real-world test results for recent theoretical findings on the generalisation performance of committees
Part of the book series: Perspectives in Neural Computing (PERSPECT.NEURAL)
Access this book
Tax calculation will be finalised at checkout
Other ways to access
About this book
Conventional applications of neural networks usually predict a single value as a function of given inputs. In forecasting, for example, a standard objective is to predict the future value of some entity of interest on the basis of a time series of past measurements or observations. Typical training schemes aim to minimise the sum of squared deviations between predicted and actual values (the 'targets'), by which, ideally, the network learns the conditional mean of the target given the input. If the underlying conditional distribution is Gaus sian or at least unimodal, this may be a satisfactory approach. However, for a multimodal distribution, the conditional mean does not capture the relevant features of the system, and the prediction performance will, in general, be very poor. This calls for a more powerful and sophisticated model, which can learn the whole conditional probability distribution. Chapter 1 demonstrates that even for a deterministic system and 'be nign' Gaussian observational noise, the conditional distribution of a future observation, conditional on a set of past observations, can become strongly skewed and multimodal. In Chapter 2, a general neural network structure for modelling conditional probability densities is derived, and it is shown that a universal approximator for this extended task requires at least two hidden layers. A training scheme is developed from a maximum likelihood approach in Chapter 3, and the performance ofthis method is demonstrated on three stochastic time series in chapters 4 and 5.
Similar content being viewed by others
Table of contents (18 chapters)
-
Front Matter
-
Back Matter
Authors and Affiliations
Accessibility Information
PDF accessibility summary
This PDF is not accessible. It is based on scanned pages and does not support features such as screen reader compatibility or descriptions for non-text content (e.g., images and graphs). However, it likely supports searchable and selectable text based on OCR (Optical Character Recognition). Users with accessibility needs may not be able to use this content effectively. Please contact us at through this accessibility request webform if you require assistance or an alternative format.
Bibliographic Information
Book Title: Neural Networks for Conditional Probability Estimation
Book Subtitle: Forecasting Beyond Point Predictions
Authors: Dirk Husmeier
Series Title: Perspectives in Neural Computing
DOI: https://doi.org/10.1007/978-1-4471-0847-4
Publisher: Springer London
-
eBook Packages: Springer Book Archive
Copyright Information: Springer-Verlag London Limited 1999
Softcover ISBN: 978-1-85233-095-8Published: 22 February 1999
eBook ISBN: 978-1-4471-0847-4Published: 06 December 2012
Series ISSN: 1431-6854
Edition Number: 1
Number of Pages: XXIII, 275
Number of Illustrations: 13 b/w illustrations
Topics: Pattern Recognition, Complex Systems, Neurobiology, Statistical Physics and Dynamical Systems
