Improving Students’ Experiences of Learning Environment Using Data Mining Techniques
I.N.M. Shaharanee1, M.S.A. Bakar2, S.Z. Saad3, J.M. Jamil4
1I.N.M. Shaharanee, School of Quantitative Sciences, Universiti Utara Malaysia, Sintok, Kedah, Malaysia.
2M.S.A. Bakar, School of Computing, Universiti Utara Malaysia, Sintok, Kedah, Malaysia.
3S.Z. Saad, School of Computing, Universiti Utara Malaysia, Sintok, Kedah, Malaysia.
4J.M. Jamil, School of Quantitative Sciences, Universiti Utara Malaysia, Sintok, Kedah, Malaysia.
Manuscript received on 25 March 2019 | Revised Manuscript received on 04 April 2019 | Manuscript Published on 27 April 2019 | PP: 202-207 | Volume-7 Issue-6S2 April 2019 | Retrieval Number: F10300476S219/2019©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: E-Course Evaluation System (e-CEvas) is an online academic course evaluation based in Universiti Utara Malaysia (UUM). This assessment gives opportunities for students in order to give feedback regarding the process of teaching and learning that they go through during each semester. Classifying lecturers’ performances based on this course evaluation data is an intriguing yet challenging problem for any academic institution. This course evaluation offers lecturers/instructors to understand their strengths as well as the weaknesses in teaching and learning processes. Referring to the course evaluation data, there are many factors and variables involve in evaluating teaching and learning delivery processes. This research work proposed a decision tree model to classify lecturers’ performances based on the course evaluation data to improve student learning experiences. We compared different data partitioning strategy and several measures of impurity to improve the classification accuracy while maintain satisfactory overall classification performance. The result indicates that, decision tree with two split combining with variable selection technique achieved the best classification performance with a 96.33% overall accuracy. We also identified the most important factors for accurate classify of lecturer performance in teaching and learning process that can improve student learning experiences. Application of these models has the potential to accurately classify under performing lecturer and help them to improve their teaching and learning thus improving student experiences in learning.
Keywords: Students ‘Experiences, Classification Model, Learning Environment, and Teacher Performances.
Scope of the Article: Data Mining