Oksanich I. Methods and models of evaluation and improving the effectiveness of problem-oriented user interface of information retrieval system.

The dissertation is devoted to the ways and methods of problem solving efficiency user interface (UI) of information retrieval systems (in particular by reducing complexity of requests for information resources). It is based on general and applied methods of probability theory and mathematical statistics, algebra of logic and mathematical methods and simulation. A model of decomposition of GOMS-KLM mental operators for the tasks of data writing and visual identification was proposed and experimentally investigated. There are constructed and studied mathematical logic-probabilistic models of "step-by-step" and "taking aim" prompts, dependencies between variants, important parameters and characteristics of UI, including the probability distribution function of the words of basic vocabulary reference. A simulation model of UI that confirm the adequacy of mathematical model was built and tested. The technology of kvaziarbitrary user queries to a database in natural-language environment using standard CASE - tools is proposed.