A Soft Set Approach for Clustering Student Assesment Datasets

Educational data mining has been studied extensively as it provides useful information for educators to make more accurate decisions concerning their students, and to adapt their teaching strategies accordingly. Data clustering as one of data mining techniques can be considered as an alternative method for educational data mining. In this paper, a data clustering technique based on soft set theory is presented. The Maximum Degree of Domination in soft set theory (MDDS) is proposed and further applied to select the best attribute in educational data clustering. To find meaningful clusters from a dataset, clustering attribute selection is conducted so that attributes within the clusters made will have a high correlation or high interdependence to each other while the attributes in other clusters are less correlated or more independent. The datasets are taken from a survey from a number of courses at the Information Engineering and the Architecture Departments of the University Technology of Yogyakarta, Indonesia. The evaluation criteria uses score range from 0 to 100. Student name, age, race, and attendance are not required in this assessment. In the results, we show how to determine the dominant attributes of a set of attributes of an assessment list by using the proposed technique. The results obtained can potentially contribute to give a recommendation in awarding the final grade of a course more quickly and accurately