ОЦЕНКА ДОСТОВЕРНОСТИ ПРЕОБРАЗОВАНИЯ ИНФОРМАЦИИ В ИНФОРМАЦИОННЫХ СИСТЕМАХ

Abstract

Objective: To find the method of data processing validation in the process of data systems engineering (IS), that provides for reliability of information at different stages of data processing (keying, reading different types of transfer support, storing, channeling, processing, presenting results to the user, etc.). Methods: The methods of statistical theory of reliability and the theory of information coding were applied, as the reliability of control options (by a human-operator, by means of hardware device, via program, etc.) significantly differs which leads to efficiency changes of one and the same method, depending on its concrete implementation. Measures on reducing a number of occurring distortions were taken in addition to error control methods. Commonly used parameters of the given methods prevent from the unified position assessment of the simultaneous application effect of different control methods and measures, aimed at reducing the number of errors, not related to control, taking into account their mutual influence, reliability and implementation structure. Results: The task of using different methods of error control (protective coding, checking sums application, duplication of operations with results comparison, data visual inspection, acceptable range check when entering some data, etc.) was solved, as well as the methods not related to control (the selection of input values from a pre-arranged list, errorfree performance of the staff incentive, etc.). In the process of IS engineering it became possible to assess the effect from application of a separate method and a group of different methods of reliability improvement at different stages of data processing. Practical importance: The considered approach makes it possible to assess detectivity of separate and the complex of control methods at the stages of data conversion in the process of IS engineering, as well as to consider from unified positions the impact of measures, not related to control, on information integrity.