Prediction Model of School Readiness

Abstract

Studying the school readiness is an interesting domain that has attracted the attention of the public and private sectors in education. Researchers have developed some techniques for assessing the readiness of preschool kids to start school. Here we benefit from an integrated approach which combines Data Mining (DM) and social network analysis towards a robust solution. The main objective of this study is to explore the socio-demographic variables (age, gender, parents' education, parents' work status, and class and neighbourhood peers influence), achievement data (Arithmetic Readiness, Cognitive Development, Language Development, Phonological Awareness), and data that may impact school readiness. To achieve this, we propose to apply DM techniques to predict school readiness. Real data on 306 preschool children was used from four different elementary schools: (1) Life school for Creativity and Excellence a private school located in Ramah village, (2) Sisters of Saint Joseph missionary school located in Nazareth, (3) Franciscan missionary school located in Nazareth and (4) Al-Razi public school located in Nazareth, and white-box classification methods, such as induction rules were employed. Experiments attempt to improve their accuracy for predicting which children might fail or dropout by first, using all the available attributes; next, selecting the best attributes; and finally, rebalancing data and using cost sensitive classification. The outcomes have been compared and the models with the best results are shown.