Abstract:

The paper summarizes the experience obtained during the last decade from usage of developed system IKAS. The IKAS is the adaptive intelligent system for knowledge assessment and self-assessment based on concept maps (CMs). The kernel of the IKAS is the intelligent knowledge assessment agent which is implemented as a multiagent system. The latter consists of the agent-expert, the communication agent, the knowledge evaluation agent, and the interaction registering agent. The knowledge evaluation agent compares a teacher’s and a learner’s CMs on the basis of graph patterns taking or not into account semantics of links (depending on the given tasks) and assigns score for a submitted solution. The paper focuses on architecture and general principles of functioning of the IKAS, the variety of provided tasks (the degree of task difficulty is changed by the system), teacher’s and students’ support and system’s feedback as well as scoring and adaptation mechanisms. Lessons learnt from practical testing of the IKAS in different study courses are discussed both from teachers and students point of view. First, the advantages of knowledge assessment based on CMs are at least twofold – it is possible to assess students’ knowledge structure and to automate the assessment process thus reducing the workload of teachers. At the same time from the usage of the IKAS it is concluded that CM based knowledge assessment is very well suitable for self-assessment, but it is not fully acceptable for giving the final marks to students for their knowledge in the subject. Results of practical use of the IKAS are presented.

Keywords:

Concept map, knowledge assessment, multiagent system.