IRO Journals
Journal of Artificial Intelligence and Capsule Networks

Predictive Al for Enhancing Clinical Pathways in Children with Learning Disabilities using XGBoost

Open Access
Published: 15 October, 2026
Pages: 361-373
  • Sujitha V. , Sujitha V.
    Head of the Department, Department of Computer Science and Engineering, NPR College of Engineering and Technology, Natham, Dindigul, India
    Head of the Department, Department of Computer Science and Engineering, NPR College of Engineering and Technology, Natham, Dindigul, India
  • Kishore P. , Kishore P.
    Student, Department of Computer Science and Engineering, NPR College of Engineering and Technology, Natham, Dindigul, India
    Student, Department of Computer Science and Engineering, NPR College of Engineering and Technology, Natham, Dindigul, India
  • Nevin Nehal R.S. , Nevin Nehal R.S.
    Student, Department of Computer Science and Engineering, NPR College of Engineering and Technology, Natham, Dindigul, India
    Student, Department of Computer Science and Engineering, NPR College of Engineering and Technology, Natham, Dindigul, India
  • Afzal Khan M. Afzal Khan M.
    Student, Department of Computer Science and Engineering, NPR College of Engineering and Technology, Natham, Dindigul, India
    Student, Department of Computer Science and Engineering, NPR College of Engineering and Technology, Natham, Dindigul, India
view PDF
view PDF

How to Cite

V., Sujitha, Kishore P., Nevin Nehal R.S., and Afzal Khan M. 2026. “Predictive Al for Enhancing Clinical Pathways in Children With Learning Disabilities Using XGBoost”. Journal of Artificial Intelligence and Capsule Networks 8 (4): 361-73. https://doi.org/10.36548/jaicn.2026.4.005.
ACM ACS APA ABNT Chicago Harvard IEEE MLA Turabian Vancouver AMA

Download Citation

Endnote/Zotero/Mendeley (RIS) BibTeX

Keywords

Classification
Dyslexia Screening
Random Forest
XGBoost
Feature Representation
Ensemble Learning
Binary Classification

Abstract

Dyslexia screening in children involves computing techniques to analyse the learning patterns from the structured data generated by behaviour and language assessments. This research work proposes a machine learning technique to classify dyslexic cases by implementing Random Forest and XGBoost on a publicly available dataset consisting of 3,644 observations and 196 features extracted from demographics and language tasks. Steps in data pre-processing involves the missing observations handling, categorical data encoding, non-predictive columns elimination, and stratified splitting of training and testing datasets. Hyperparameter tuning is performed via stratified cross-validation, followed by independent testing of both ensemble models. Random Forest attained 95.1% accuracy with 94.3% precision, 93.8% recall, 94.0% F1-Score, and 0.97 ROC-AUC score, whereas XGBoost gave an accuracy of 91.8% and ROC-AUC of 0.94. Flask-based prediction model was created for providing individualized screening predictions. The platform shows how ensemble learning could be used in computational dyslexia screening and serves as an illustrative example of the decision support tool for preliminary screening.

References

  1. Poornappriya, T. S., and R. Gopinath. ”Application of Machine Learning Techniques for Improving Learning Disabilities.” International Journal of Electrical Engineering and Technology (IJEET), 2020, vol. 11, no. 10: 392-402.
  2. Chen, Tianqi, and Carlos Guestrin. ”Xgboost: A Scalable Tree Boosting System.” In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2016, 785-794.
  3. Seshadri, NP Guhan, Sneha Agrawal, Bikesh Kumar Singh, B. Geethanjali, V. Mahesh, and Ram Bilas Pachori. ”EEG based Classification of Children with Learning Disabilities Using Shallow and Deep Neural Network.” Biomedical Signal Processing and Control, 2023, vol. 82: 104553.
  4. JothiPrabha, A., R. Bhargavi, and BV Deepa Rani. ”Prediction of Dyslexia Severity Levels from Fixation and Saccadic Eye Movement Using Machine Learning.” Biomedical Signal Processing and Control, 2023, vol. 79: 104094.
  5. Nerušil, Boris, Jaroslav Polec, Juraj Škunda, and Juraj Kačur. ”Eye Tracking-Based Dyslexia Detection Using a Holistic Approach.” Scientific Reports, 2021, vol. 11, no. 1: 15687.
  6. Sbiaa, Fatma, Sonia Kotel, Rania Mghirbi, and Ahmed Ghazi Blaeich. ”Revolutionizing Dyslexia Diagnosis: An Intelligent Model Featuring Machine Learning and Fuzzyfication.” Procedia Computer Science, 2024, vol. 246: 3624-3633.
  7. Liu, Fulin, Hongan Wang, Xin Zhao, and Dongchuan Yu. ”Small-world Properties of Eye-Movement Time Series Assisted in Identifying Children at High Risk for Dyslexia.” Biomedical Signal Processing and Control, 2024, vol. 93: 106148.
  8. Nagarajan, Harshitha, Vishnu Sai Inakollu, Punitha Vancha, and J. Amudha. ”Detection of Reading Impairment from Eye-Gaze Behaviour Using Reinforcement Learning.” Procedia Computer Science, 2023, vol. 218: 2734-2743.
  9. Ding, Ning, Peng Peng, Jiuqing Tang, Yiran Ding, and Jingjing Zhao. ”An Investigation of Phonological Predictors in Chinese Developmental Dyslexia Using a Machine Learning Approach.” Reading and Writing, 2025, vol. 38, no. 10: 2849-2873.
  10. Odoom, Samuel, Eric Opoku Osei, Enock Quansah Effah, Victoria Boafo, Seyram Dusu, and Anthony Kweku Appiah. ”A 3-tier Machine Learning Framework for Early Detection of Learning Difficulties In Basic School Settings.” Discover Global Society, 2026, vol. 4, no. 1: 60.
  11. L. Rello, R. Baeza-Yates, A. Ali, J. P. Bigham, and M. Serra, “Predicting Risk of Dyslexia with an Online Gamified Test,” PLOS ONE, 2020, vol. 15, no. 12: e0241687. Available: https://www.kaggle.com/datasets/luzrello/dyslexia