Analysis of Ramer-Douglas-Peucker Algorithm As a Discretization Method


GÖZ F. , MUTLU A. , AKBULUT O.

26th IEEE Signal Processing and Communications Applications Conference (SIU), İzmir, Türkiye, 2 - 05 Mayıs 2018 identifier identifier

Özet

Discretization is the process of converting continuous values into discrete values. It is crucial for several machine learning and data mining algorithms as certain algorithms work only on discrete values. In this study we investigate applicability of Ramer-Douglas-Peucker (RDP) algorithm as a discretization method. Experimental results demonstrate that RDP-based discretization achieves similar or better classification accuracy compared to equal width, equal frequency, Zeta and 1R.