Integrating Mental Models into Intelligent Tutoring Systems for Solving Random Sampling Word Problems

Mental models have been deeply concerned by researchers and practitioners since it plays a crucial role in problem-solving. In this study, we proposed mental models for random sampling word problems. On basis of the proposed mental models, we further performed a mental simulation with a rule-based mental reasoning method and a template-based approach to generate human-style explanations. And then we established an experiment on a prototype intelligent tutoring system. Experimental results showed that: the proposed mental models were available for most random sampling word problems; and a solution with human-style explanations to the problem encoded in mental models was produced effectively by means of mental reasoning. Based on these results, we concluded that our mental models can be used to encode essential information for solving random sampling word problems effectively, and it is profitable to take mental models into account when developing intelligent tutoring systems.