A Federated Learning Approach for Personalized Online Course Recommendation Using Hybrid Recommender Strategies
Bibliographic record
Abstract
The rapid pace of growth in online learning environments has made it more difficult for learners to find courses that are tailored to their individual learning needs and goals. Typically, recommender systems rely on centralized data processing, which raises considerable concerns around user privacy and the heterogeneity of user data. This research presents a hybrid recommender system with Federated Learning (FL) characteristics to provide personalized recommendations to users while protecting their data. The recommender system creates four primary recommendation frameworks: Content-Based Filtering (CBF); Popularity-Based; Collaborative Filtering (simulated); and a novel Federated Learning approach that hybridizes on the three other recommended/first ranked approach, whose performance is evaluated with a graphical user interface (GUI) built using Tkinter that provides real-time recommendations to chosen users and comparisons amongst the four recommended approaches using multiple performance measures. The assessment of the systems' efficacy used seven typical evaluation metrics: Hit Rate, Precision@5, Recall@5, F1@5, Mean Reciprocal Rank (MRR), Normalized Discounted Cumulative Gain (NDCG@5), and Coverage. The bar charts demonstrated that while the Popularity-Based recommendation approach had the highest hit rate respectively and highest precision, it had very low coverage of recommending courses to learners. On the other hand, the Federated Learning-based approach produced a comparable hit rate and precision, while also achieving comparable F1 scores and coverage (86 %). This study emphasized the benefits and practicality of using Federated Learning to build contextually scalable, safe, and effective course recommender systems, and further emphasized the benefits of combining or hybridizing the two approaches for maximum robustness and quality of recommendations, especially for the educationallyfocused digital sphere.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".