Machine Learning Approach to Predict Basic Literacy Skills of Elementary School Students

Early identification of elementary students' literacy levels is vital for many reasons. With the availability of an extensive collection of data and advanced Machine learning (ML) algorithms, students' literacy levels can be predicted to ensure they are meeting the benchmark level. This manuscript presents the findings of such an investigation to predict end-of-the-year literacy levels of students at the elementary level. Using a dataset collected about students' related information and their academic scores (DIBELS), five machine learning models are constructed to classify whether students meet standards. Based on the experimental findings, the hyper-parameterized Random Forest model shows a significant performance with 81% of recall value in identifying a student at the risk of meeting the benchmark level.