N-umwali/rwanda-academic-pathway-recommender

# Rwanda Academic Pathway and Bridge Course Recommendation System ## Project Overview The Rwanda Academic Pathway and Bridge Course Recommendation System is a Streamlit-based decision-support prototype developed for Rwandan General Education and TVET learners. It uses a trained Support Vector Machine model to rank broad academic fields, then applies pathway-aware guidance to refine the result into a specific program direction, bridge-course preparation, and an alternative pathway. The system is advisory. It does not replace official university admission requirements, institutional guidance, or professional academic counselling. ## Main Functions - Collects a structured learner profile for General Education and TVET routes. - Predicts one of 16 broad academic program categories using a trained LinearSVC pipeline. - Refines the broad prediction into a learner-facing program recommendation. - Keeps TVET recommendations grounded in the learner's current trade while recording broader interests as possible transition goals. - Maps the recommendation to bridge courses and an alternative pathway. - Uses SHAP values from the fitted linear model to identify the profile features that most influenced the broad-field prediction. - Generates a downloadable guidance report. - Collects prototype feedback on usefulness, clarity, and usability. ## Recommendation Architecture ```text Learner profile ↓ Input validation and model-vocabulary preparation ↓ LinearSVC ranks 16 broad academic fields ↓ Pathway and TVET-alignment refinement ↓ Specific program recommendation ↓ Bridge-course and alternative-pathway mapping ↓ SHAP-supported learner-friendly explanation ↓ Human review and verification of official requirements ``` The machine-learning prediction is produced first. The guidance layer then converts the broad model output into a practical recommendation and applies pathway-specific safeguards. SHAP explains the trained model's broad-category decision; it does not claim that th …