Title: Statistical characteristics for multi-dividing ontology algorithm in AUC criterion setting

Authors: Yun Gao; Linli Zhu; Wei Gao

Addresses: Editorial Department of Yunnan Normal University, Kunming 650092, China ' School of Computer Engineering, Jiangsu University of Technology, Changzhou 213001, China ' School of Information, Yunnan Normal University, Kunming 650500, China

Abstract: The architecture model of ontology has been extensively applied in many fields including computer science and information technology. The key purpose of using ontology is to compute the similarity among the vertices in the ontology graph. The scientific method of multi-dividing technology can accurately draw the ontology graph into real line and measure the similarity among different vertices on the basis of the difference of equivalent real numbers. This paper explores the theoretical problem of multi-dividing ontology algorithm in AUC criterion setting. The approximation characters and the generalised error of best ontology score function among piece constant ontology score functions are presented. Also, the rate bound for the best ontology function under penalised empirical AUC criterion is given. The results we got in the paper illustrate the promising application prospects for multi-dividing ontology algorithm.

Keywords: ontologies; similarity measures; AUC criterion; piecewise constant ontology score function; ROC curve; penalty terms; multi-dividing ontology.

DOI: 10.1504/IJCI.2016.077100

International Journal of Collaborative Intelligence, 2016 Vol.1 No.3, pp.178 - 188

Received: 22 Jan 2015
Accepted: 25 Feb 2015

Published online: 21 Jun 2016 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article