Koliada A. Models and methods of information search in scientometric databases

The thesis is devoted to the problem of creating information technology for extracting metadata from scientometric publications database through the web interface. The definition Scientometrics database shows the most common characteristic of them, and methods of use of these databases. The model of extracting information from poorly structured web pages and automate the process of extracting the model of many scientometric databases. Also improved method of extracting information from dynamic web pages that require code execution on the user side. The process of thematic modeling, and applied-latent semantic analysis and latent Dirichlet allocation of names to the list of publications seized to allocate them to close in content topics. The software automation system metadata extraction publications from the most common scientometric databases with graphic user interface to manage search publications viewing and analysis.