Models and procedures for interval evaluating the results of control of knowledge in computer systems testing of Navy

Abstract

To implement effective military and professional training of Navy specialists, a corresponding educational and material base is needed. As a result of the reduction in the 1990s in the branches of the military-industrial complex developing weapons and equipment for the Navy, the latest models of this technology are now produced not in batches, but in individual copies. The question of the production of training and training samples is not worth it at all. Under these conditions, only virtual analogues of military equipment and weapons, developed by means of information technology, i.e., training and training systems (TOS), can be considered as the only means capable of providing military-professional training. At the modern level of the development of information technologies, testing is the only universal technical means of monitoring the knowledge of students. Procedures for knowledge control in modern computer testing systems do not meet the requirements for them according to the following characteristics: 1) the absence of the possibility of evaluating the error of the test results; 2) the absence of the possibility of stopping testing when the specified reliability of its results is achieved. In order to effectively implement the means of operational criteria-based pedagogical control of knowledge in the process of training specialists of the Navy and to enable joint analysis and processing of evaluations of learning outcomes, it is advisable to implement the following practical recommendations: 1. Formulating the teacher's preferences system regarding the quality of trainee training and the teacher's preferences system in relation to The significance of single test tasks in the test should be considered as the most important The essential steps in preparing a test for practical use. 2. The teacher who first enters the input of his preference systems should check their actual compliance on a sample of 5-10 such test results that cover the full range of possible outcomes (from fully faithful to completely incorrect). The presence of the developed and presented models provides the possibility of setting and solving the task of developing models and procedures for managing the process of criterion-oriented testing, ensuring the efficiency of this process.