A Posteriori Corrections to Classification Methods
A posteriori corrections are computational inexpensive and may improve accuracy, confidence, sensitivity or specificity of the model, or correct for the differences between a priori training and real (test) class distributions. Such corrections are applicable to neural and any other classification models.
KeywordsCost Function Rejection Rate Class Distribution True Class Logical Rule
Unable to display preview. Download preview PDF.
- 2.Saerens M, Lattinne P, Decaestecker C. (2002) Adjusting the outputs of a classifier to new a priori probabilities: a simple procedure. Neural Computations (in press).Google Scholar