Data-Driven Weak-Subject Identification and Multi-Technique Personalization for Adaptive E-Learning: An XGBoost, TF-IDF, and Constraint Programming Framework

Author(s)	Mr. Arun Karthik V, Dr. K. Karthikayani
Country	India
Abstract	A data-driven personalization framework for K-12 e-learning is presented that replaces manually declared subject preferences with an objective, comparative analysis of a student's grades across all subjects. The system integrates five distinct algorithmic components into a single deployed pipeline: an Extreme Gradient Boosting (XGBoost) regressor that predicts a student's next assessment score specifically in the subject identified as weakest; a Term Frequency–Inverse Document Frequency (TF-IDF) and cosine-similarity content recommender that ranks learning resources by relevance to that weak subject and the student's learning style; a transparent rule-based engine that converts the prediction into a daily study plan; a Constraint Programming–Satisfiability (CP-SAT) solver from Google OR-Tools that packs recommended resources into an optimised weekly schedule under daily time and session constraints; and a rule-based conversational tutor and faculty-facing analytics dashboard.
Keywords	Personalized Learning, Educational Data Mining, XGBoost, TF-IDF, Cosine Similarity, Constraint Programming, CP-SAT, Adaptive Learning, Intelligent Tutoring Systems, Streamlit.
Field	Computer > Artificial Intelligence / Simulation / Virtual Reality
Published In	Volume 8, Issue 5, September-October 2026
Published On	2026-10-07

About IJFMR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9687-828-838	editor@ijfmr.com

International Journal For Multidisciplinary Research