Using Dempster-Shafer theory and ontology in XML latent information retrieval

In this paper we describe an XML latent information retrieval model based on Dempster-Shafer theory and ontology that tries to find the latent information in XML documents in response to query. In this model, we utilize ontology to get the related instances in XML elements, propose two algorithms that use Dempster-Shafer theory of evidence to combine the related instances, which are in the same class within the ontology. Furthermore, we present the model´s application in the detection of personal and enterprises´ sensitive information in e-government domain. Experiments prove that the proposed model can retrieve latent information and has a higher performance than traditional methods.