Pre-assessment and learning recommendation mechanism for a multi-agent system

Abstract

Diagnostic assessment is a vital and effective strategy in any teaching-learning process such that it provides a pre-learning assessment of the learners state of knowing with regard to a given knowledge concept. Current intelligent learning systems still do not integrate effective techniques for evaluating prior knowledge that can be used effectively to diagnose gaps that will inhibit future learning and for making recommendations for learning and tutoring to fill them. In this paper, we present a mechanism for pre-assessment of previous learning upon which the recommendation for a new or appropriate learning level is based. Our approach is based on message passing procedure between agents in a multi-agent system. We have tested the pre-assessment technique with a prototype based on the Jason AgentSpeak language, and using learning materials from a structured query language (SQL) revision module.