docNumModel

Centre for Vision, Speech & Signal Processing

Numerical Models Class Structure and Usage

NumModelC is a base handle class for numerical models of the form Y=f(X) where Y and X are n and m dimensional vectors respectively. It differs from NumFuncC in that the parameters of the model can be fitted using sample data in a training set. Training the model is achieved with a call to Fit supplied with a NumVVDataSetC sample set.

The first time that a model is fitted with data it is automatically initialised. Depending on the model fitting algorithm, it is not explicitly initialised again until the Initialise function is called again. This means that for things like a neural network the training can be performed incrementally.

Author: Robert Crida, Created: 15/6/1998, Generated by DocCreate: March 20, 2001