Use of thesauruses for search of complex information objects on Web on the basis of ontologies

Abstract

An ontological model of interaction between objects and subjects of the Web semantic search is proposed, its basic elements are characterized, methods for its replenishment and use for filtering information that is pertinent  to the personalized needs of users are considered. The types of relations between the instances and classes of this model and their characteristics that can have an influence on the time complexity of processing knowledge presented on the basis of this model, are analyzed. One of the important elements of the proposed model are thesauri that represent  knowledge regarding tasks for which users are looking for information, and about information resources where such information can be contained. The expediency of using particular cases of ontology – thesauruses – to find semantically similar information objects is substantiated. The types of thesauruses that are used for semantic search are considered, methods for their replenishment and utilization are proposed, and their characteristics are analyzed. In this paper we proposes an algorithm for the automated construction of simple thesaurus, which is formed on the base of the domain ontology  and a natural language description of the user’s problem, and methods for generating composite thesauri that are pertinent to new user tasks which are based on the set of simple thesauruses previously built by user. The expressiveness and computational complexity of the proposed methods which depend on the domain ontology characteristics and on the size of the problem description are estimated. Methods of use of the semantically marked Wiki resources as a source of knowledge for constructing ontologies of subject areas and associated typical information objects are considered.

Problems in programming 2019; 4: 28-52