Abstract
So far, we have not addressed the question of how the components of the feature vector X are obtained. In general, these components are based on d measurements of the object to be classified. How many measurements should be made? What should these measurements be? We study these questions in this chapter. General recipes are hard to give as the answers depend on the specific problem. However, there are some rules of thumb that should be followed. One such rule is that noisy measurements, that is, components that are independent of Y, should be avoided. Also, adding a component that is a function of other components is useless. A necessary and sufficient condition for measurements providing additional information is given in Problem 32.1.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 1996 Springer Science+Business Media New York
About this chapter
Cite this chapter
Devroye, L., Györfi, L., Lugosi, G. (1996). Feature Extraction. In: A Probabilistic Theory of Pattern Recognition. Stochastic Modelling and Applied Probability, vol 31. Springer, New York, NY. https://doi.org/10.1007/978-1-4612-0711-5_32
Download citation
DOI: https://doi.org/10.1007/978-1-4612-0711-5_32
Publisher Name: Springer, New York, NY
Print ISBN: 978-1-4612-6877-2
Online ISBN: 978-1-4612-0711-5
eBook Packages: Springer Book Archive