Analysis for Semi-supervised Learning Ontology Algorithm
Wei Dong Gao · 2015
Ontology similarity measure and ontology mapping are widely used in knowledge representation and information processing. One method to get ontology algorithm is using graph Laplacian semi-supervised learning method, all the vertices of the ontology graph are mapped into real numbers. Then ontology similarity measure algorithm is obtained by comparing the difference of their corresponding values. In this paper, the stability of ontology algorithms is studied by adopting a strategy which adjusts the sample set by deleting one element from it. The generalized bound on such leave-one-out stability is given.