Testing and Validation

Abstract

Some authors claim that one of the main challenges for researchers seeking to implement and to validate a new ID method is to assess and compare its performance with other available approaches [301]. Benchmarking an IDS is a common way of establishing its effectiveness, although this task is neither easy nor clear-cut [16, 302]. It is even more difficult when dealing with visualisation-oriented IDSs as no specific comparative techniques exist, unlike those for other approaches such as the ROC (Receiver Operating Characteristic) analysis [303].