Introduction to error estimation in pattern recognition

The training objectives of the learning object are: 1) To compute the theoretical error of a classifier; 2) To compute the Bayes error: 3) To estimate the error of a classifier by resubstitution; and 4) To estimate the error of a classifier by holdout and add a 95% confidence interval. They are covered at a basic level with the help of simple examples. Juan Císcar, A.; Sanchis Navarro, JA.; Civera Saiz, J. (2019). Introduction to error estimation in pattern recognition. http://hdl.handle.net/10251/121153