Abstract:

Automated knowledge assessment is an important component of e-Learning systems to ensure effective management of the learning process. As a rule, in modern e-learning systems, the evaluation of the learner's knowledge is carried out through testing - through choosing the right variant from the set of presented answers. It is obvious, that multiple-choice tests can’t provide full evaluation of learner’s knowledge, thereafter, they do not provide a flexible and qualitative management of training. Thus, it is increasingly important problem to implement an e-Assessment system featuring support of answers that are freely formulated in natural language. An analysis of the most recent publications indicates that automated assessment of natural language QA-texts remains weakly developed in modern e-Learning systems. In this paper an experimental prototype of e-Assessment system featuring automated evaluation of natural language answers is presented. The theoretical model presented by D. Suleymanov is used for develop of answer processor component. An experiment on implemented prototype of the system was conducted to acquire the data needed for further development of algorithm for analysis of natural language QA-texts within the e-Learning system. Experiment results and conclusions are presented.

Keywords:

Natural language processing, NLP, e-Assessment, automated assessment, e-learning, automatic answer grading.