October 2008
Intermediate to advanced
984 pages
30h 43m
English
This chapter is the last one related to supervised learning, and it is intended to serve three purposes. The first sections focus on the last stage of the design procedure of a classification system. In other words, we assume that an optimal classifier has been designed, based on a selected set of training feature vectors. Our goal now is to evaluate its performance with respect to the probability of classification error associated with the designed system. To this end, methodologies will be developed for the estimation of the classification error probability, using the available, hence finite, set of data. Once the estimated error is considered satisfactory, full evaluation of ...
Read now
Unlock full access