Voborník P.: Optimization of the Method for Selection of Questions into the Repeated Electronic Test for the Purpose of Teaching and Evaluated Testing. In: DIVAI 2016 – The 11th international scientific conference on Distance Learning in Applied Informatics. Štúrovo, Slovakia: Wolters Kluwer, 2016, pp. 445–455. ISBN 978-80-7552-249-8, ISSN 2464-7470, e-ISSN 2464-7489.

Abstract

Optimization of selection of questions into the test from a larger set of issues is a problematic topic, often solved in test systems, which is usually processed only by simple random sampling. An original proposal of a sophisticated selection method customized to the test applications will be presented in this article in detail. The principle is based on a modern version of the theory of forgetting. This theory is modified by the possibility of determining the period of testing and it is also complemented by the knowledge factor. The probability of selecting each of the questions into the test are then determined from the combination of these components according to the principles of genetic algorithms. Thanks to this method, tested users are presented more often with such questions the topics of which these users don't have sufficiently fixed in their memory, and those they should revive again. In evaluated testing, this method can check with students whether they already have learned the theme, which they did not known the last time. This model can be an inspiration not only for the implementation of future testing applications but also for further development of the theory of choice functions, ether in this or an entirely different science branch.

Keywords

Test, forgetting curve, spaced repetition, choice function, Ebbinghaus, genetic algorithm.

Dostupné on-line.