Identifying the contextual factors related to the reading performance of Moroccan fourth-grade students from a Machine Learning-based Approach.

Based on the 2016 Progress in International Reading Literacy Research (PIRLS) data, this study determined the contextual characteristics that distinguished Moroccan fourth-grade kids with high- and low-achieving reading abilities. The PIRLS 2016 public dataset included 5446 Moroccan pupils, with 1545 scoring well and 3901 doing poorly on the reading portion of the exam. The objective of the present research was to determine the ideal combination of significant contextual elements that jointly affect fourth-grade children's reading ability. Machine learning technique such as gradient boosting was used to find the most critical contextual variables that might distinguish between the two groups of students. The findings demonstrate that 12 significant contextual variables from GradientBoosting, selected from 486 contextual variables at the levels of the student, teacher, and school, Identifying high- and low-performing students collectively have consequences for teaching and learning in the future about the reading ability of primary school students.