Neural Networks for Conditional Probability Estimation

Forecasting Beyond Point Predictions

By:

Dirk Husmeier

Format:

Paperback / softback

Series:

Perspectives in Neural Computing

Show Hide Adult Content

$134.99

RRP:: ~~$187.95~~ save $52.96

Available from supplier

Usually ships in 3-4 weeks

0 in Trolley

Price Match

Wrap this item as a gift ($5.90) - add a free gift card on the next page

Buy Now, Pay Later with:

4 payments of $33.75 with Afterpay Learn more

Availability

Delivering to:

Estimated arrival:

Around 20 Jun - 2 Jul using International Courier

Description

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.

Release date Australia

February 22nd, 1999

Author

Dirk Husmeier

Audience

Professional & Vocational

Edition

Softcover reprint of the original 1st ed. 1999

Illustrations

13 Illustrations, black and white; XXIII, 275 p. 13 illus.

Pages

275

Series

Perspectives in Neural Computing

Dimensions

155x235x16

ISBN-13

9781852330958

Product ID

5263818

Customer reviews

Nobody has reviewed this product yet. You could be the first!

Write a Review

Marketplace listings

There are no Marketplace listings available for this product currently.
Already own it? Create a free listing and pay just 9% commission when it sells!

Sell Yours Here

Get free delivery, exclusive deals, and more^*

Non-Fiction Books:

Neural Networks for Conditional Probability Estimation

By:

Format:

Series:

Buy Now, Pay Later with:

Availability

Description

Customer reviews

Marketplace listings

Help & options

Filed under...

Get free delivery, exclusive deals, and more*

Buy Now, Pay Later with:

Get free delivery, exclusive deals, and more^*