This study selected OECD 35 economy members’ science, math, and reading scores and related impact factors as targets to mining the patterns and explore the main factors impact on the PISA2015 performance. The data selection was the first step; then this study applied observation clustering function with Minitab to determine the optimal clusters. The 3D scatterplot and 3D surface plot have been used to display the data structure. The dendrogram with three clusters drew by Ward linkage and Euclidean distance has a relatively high similarity level and a relatively low distance level in this study. The result reveals OECD economies in the cluster1 and cluster2 are needed to improve their students’ performance. The teaching hours per year in OECD economies has negative relationship with PISA2015 performance. While the teaching hours per year in economies can explain only 12.50% of the OECD/PISA2015 performance in the regression model. The OECD/PISA data provides an excellent databank for mining practices.
ICIC Express Letters Part B: Applications, 9(11), pp.1139-1146