Linear Statistical Analysis of Multi-dividing Ontology Algorithm ⋆

As a concept structured model, ontology has been widely used in various fields such as query expansion and image retrieval. The core trick of ontology applications is to calculate the similarity between the vertices in the ontology graph. Multi-dividing technology has been proved to be an effective approach, which maps the ontology graph into real line and determines the similarity between the different vertices according to the difference of real numbers they correspond to. In this paper, we study the multi-dividing ontology algorithm from a theoretical view. It is highlighted that empirical multi-dividing ontology model can be expressed as conditional linear statistical, and an approximation result is achieved based on projection method. Moreover, we determine the bound of algorithm error for W -ontology performance criteria, which we propose for multi-dividing ontology setting.