Title: Empowering education by harnessing machine learning for predicting student readability score

Authors: Huixian Li; Long Li

Addresses: Basic Teaching Department, Hebei Vocational University of Industry and Technology, Shijiazhuang, Hebei, China ' Basic Teaching Department, Hebei Vocational University of Industry and Technology, Shijiazhuang, Hebei, China

Abstract: Readability refers to the ease or difficulty with which a reader can understand a given text. In this study, machine learning techniques were used to predict students' reading scores, with Random Forest Classification (RFC) demonstrating promising performance. The effectiveness of RFC, Sooty Tern Optimisation Algorithm (RFST) and Gold Rush Optimiser (RFGR) was evaluated using 1000 English texts based on accuracy, precision and F1-score. Results show that RFST outperformed RFC and RFGR, achieving an average training accuracy of 0.968 across various reading levels. This superior performance is attributed to its enhanced ability to capture complex linguistic patterns. In contrast, RFC and RFGR achieved accuracies of 0.932 and 0.935, respectively. The study also emphasises the critical role of attribute selection, particularly lexical variables, in improving predictive accuracy. These findings contribute to the advancement of readability assessment and support the development of more accurate readability prediction frameworks.

Keywords: readability; reading score; random forest classification; sooty tern optimisation algorithm; gold rush optimiser.

DOI: 10.1504/IJIPT.2026.156308

International Journal of Internet Protocol Technology, 2026 Vol.19 No.3, pp.140 - 155

Received: 24 Apr 2025
Accepted: 11 Jun 2025
Published online: 10 Sep 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article