Site Map: Structured Outline

Hop to Top of Documentation Tree
Leap to Chapter-Start
Move Up a Page


FEATURE SPACE METHODS


Based on defining an m dimensional space where each of the k cases is defined as a point.

In a simple application m=n, and the x1, x2, x3, .....xn are the dimensions. In a more complicated applications the parameters are massaged into a different set of parameters.

Feature space computations are based on the distance notion of a metric space.

Methods:

  • Extended Curve Fitting
  • Cluster Analysis
  • Samid's Distance-Weight Computation

  • All these methods suffer from a high injection of arbitrariness owing to the need to establish a relative scale of the m dimensions.


















    Comment? eMail: teach_me@classexpress.com








    (c) Class Express LLC Contact Us