Fuzzy dialog analysis in reciprocal learning

The dialog behavior between two learning companions in RTS (Reciprocal Tutoring System) is analyzed through fuzzy set theory. RTS is a collaborative learning system with two learning companions playing two roles: tutor and tutee. By communication, these two learners share their ideas and solve programming exercises cooperatively. Each linguistic sentence of their talks belongs to a fuzzy set that indicates different-kinds of dialog style. Through processing the experimental data, the fuzzy dialog vectors of different styles can be calculated and all the fuzzy transition relations can also be estimated to find out all kinds of dialog path. As results, students use different dialog styles to express their intention, instead of just using one dialog mode and tutee tikes to ask tutor ″Why″ to make tutor find some correct key point. All these results can also be useful in studying student model and implementing a good virtual learning companion.