Abstract: Lifelong learning is a sustainable education, which can provide learning management as appropriate to the potential of the students. Thus, this paper is aimed to study about the factors (courses) that affect the design of learning styles in upper secondary level from lower secondary level, develop a model of education that supports the appropriate learning styles in upper secondary school, and evaluate the effectiveness of the education model. The dataset was collected from 603 lower students and 626 upper students at Phadungnaree School in Mahasarakham, Thailand. The scope of the dataset is six academic years from 2010 to 2015. The research tool consists of four machines learning, including k-means, decision tree, k-optimization, and cross-validation methods. The results showed that the number of appropriate class for school management is equal to three clusters (classrooms) while the signification factor is 21 factors in 3 groups. The model performance is very high with 82.47 % accuracy. For the future works, it can be applied in the model analysis for designing and developing applications needed for institutions.

Keywords: Educational Data Mining (EDM), Distinction Learning Style, Academic Achievement, Knowledge Management in Education

DOI: 10.1109/TIMES-iCON.2018.8621778