Ontology Algorithms Based on Mahalanobis Matrix Learning
WU Jian-zhan · Journal of Southwest University · 2015
In this paper,we present new ontology similarity measure and ontology mapping algorithms based on Mahalanobis matrix learning.Using the matrix decomposition and coordinate descent iterative algorithm,we obtain Mahalanobis matrix and distance matrix of the ontology graph,and thus determine the similarity between vertices in the ontology graph.Finally,the algorithms are applied to the biology GO ontology and computer software ontology to verify their validity by experimental data.