Decision making system to select suitable courses based on learner's learning styles using outlier analysis and classification

Students are generally forced by their parents to take different courses without knowing their inner ability and their interest. To choose suitable course it is necessary to assess their learning styles and inner ability. In this paper the learning styles of students have been assessed based on Felder and Silverman's “Index of Learning Styles questionnaire”. While collecting the data from students using ILS questionnaire, some inconsistent data may be generated due to many reasons. To build correct decision making system to select suitable courses it is needed to eliminate all such inconsistencies from the original data. The data collected from this questionnaire is a categorical data. Different outlier analysis techniques for categorical data have been used to eliminate outliers in this paper. Nearly five hundred records are collected from various engineering colleges from different branches of engineering. After eliminating the inconsistencies from data, different classification models have been built and their accuracies at different levels are compared.