Zosimov V. Models and tools of corporate web resources' intelligent data processing

The thesis, based on performed theoretical and experimental studies, is devoted to solving an important applied science problem of improving the quality of intellectual search and data processing on the Internet using the software search agents by developing a comprehensive approach that integrates new methods, models and tools into a single semantic-level system of complex web data operation in real time. Based on the analysis of information content and navigation elements of a large number of corporate web resources, their general structure and semantic profile are constructed, that is a set of semantic classes and properties that describes their content. A new approach is proposed to uniquely identify web page data snippets within a domain name, based on the use of semantic attributes of HTML blocks instead of the parent web page URL, it allows to form templates for displaying a web resource content with arbitrary arrangement of data blocks regardless of their belonging to certain pages of a web resource.