detetct outliers after classification

hi....

how can i detetct outliers after classification is done?(for exapmle decision tree learner)