SoftTutor — A XML based architecture for online questionnaire evaluation

In most of the existing Intelligent Tutoring System, the initial clustering of students is based on their previous semester marks. This paper introduces SoftTutor, an ITS for teaching complex data structures in C for 4^th semester B. Tech (Computer Science and Information Technology)students. Here instead of using their previous semester marks, a pretest is conducted by SoftTutor which includes questions in basics of computer science and a few on abstract reasoning. A XML based architecture is build for analyzing the answers. The dynamic DTD for each answer is constructed by performing apriori algorithm on the training data (collected from 60 students), to obtain the most frequent answers that are likely to be correct. Later the students are clustered based on their online marks using EM clustering algorithm. The low performing students were given additional assignments based on the clustering results so that they can improve themselves.