Modeling Primary School Student Academic Performance Using Data Mining Technique

Every time when an examination comes, the first issue discussed by all parties is student performance. That scenario describes the importance of the student performance during the examination for student. Primary school student performances have long time focused by the Ministry of Education. Thus, the primary school academic performance becomes a problem that must be solved by teachers and school administrations. They responsible to improve the students' performance from time to time for every examination held. Thus, the aim of this study is to develop a model for primary school student academic performance using data mining technique. To achieve this, a student academic profile from Sekolah Kebangsaan Kuala Teriang, Langkawi was chosen as a case study. The dataset that consist of 393 records represent the academic performance of student from standard 4 to 6 in the year 2010. The dataset consists of 6 attributes that are gender and 5 core subjects namely Pemahaman, Penulisan, Sciences, English and Mathematics which were then grouped into excellent, fair and weak group and was mined using association rules technique based on Apriori algorithm to find interesting rules which can influence student academic performance. The finding indicates that the combination attributes of Pemahaman = A, Penulisan = A, Sciences = A, English = A and Mathematics = A or B is for excellent group. The combination attributes of Pemahaman = A or B, Penulisan = A or B, Sciences = B or C, English = B and Mathematics = C or D is for fair group and the combination attributes of Pemahaman = B or C or D, Penulisan = C or D, Sciences = C or D, English = C or D or E and Mathematics = D or E is for weak group. From the finding, a prototype of the primary school academic performance was developed to help the teachers and school administration identify their student academic performance status which involves excellent, fair and weak groups.