Development of prototype of natural language answer processor for e-learning

Abstract

© Springer Nature Switzerland AG 2020. Automated knowledge control is an important component of e-Learning systems. Moreover, as an analysis of the most recent publications shows, the possibility of automated free natural language answers assessment is almost not represented in modern e-learning systems. Existing educational technologies either support only test approach to knowledge control, or when processing a natural-language answer, its semantic structure is not taken into account sufficiently for accurate assessment. This paper presents a software prototype that implements an algorithm for semantic processing of natural language answers. The basis of the algorithm is a theoretical pragmatically-oriented model proposed by D. Suleymanov, where main methodological principles are the principle of context determinism and the principle of meaning expectation. The implemented prototype was evaluated in order to verify its compliance with theoretical model and obtain the data necessary for further development of model and algorithm.