Towards an Iterative Classification Based on Concept Lattice

In this paper, we propose a generic description of the concept lattice as classifier in an iterative recognition process. We also present the development of a new structural signature adapted to noise context. The experimentation is realized on the noised symbols of GREC database [4]. Our experimentation presents a comparison with the two classical numerical classifiers that are the bayesian classifier and the nearest neighbors classifier and some comparison elements for an iterative process.

• Type: Conference
• Year: 2006
• Where: CLA
• Authors: Stéphanie Guillas, Karell Bertet, Jean-Marc Ogier
• Keywords: Bayesian Classifier, CLA 2006, CLA 2007, Classical Numerical Classifiers, Iterative Recognition Process