PULSE Index: Aspect-Based Sentiment Analysis and Topic Modeling of Student Feedback for Department Rankings in Higher Education

Academic department rankings often overlook the rich evidence in students’ open-ended comments, relying instead on questionnaires, reputation, and research metrics. This paper introduces the PULSE Index (Perceived University Learning & Student Experience), a 0 - 100 composite score that converts student comments into department-level indicators using a transparent AI/NLP pipeline. The method is student-centered by design. Stage 1 defines a set of student-valued categories (e.g., teaching, infrastructure, services) and elicits their weights via a multi-criteria procedure. Stage 2 collects and de-identifies comments, assigns each comment to categories with aspect-based classification, and estimates neutral-aware sentiment using modern embeddings. Topic models (LDA/BERTopic) provide an empirical check on the category map and surface emergent themes. For each module and category, we compute conditional mean sentiment, coverage and intensity; apply empirical-Bayes shrinkage to stabilize small samples; and, optionally, fold salience and strength into a representativeness factor. Stage 3 normalizes indicators within peer groups, forms module composites via the elicited weights, and aggregates modules to departments using reliability- and volume-aware weights, before scaling to the 0 - 100 PULSE Index. Robustness is assessed through sensitivity analysis (weights, normalization), bootstrap confidence intervals and rank-stability; fairness diagnostics examine subgroup patterns (e.g., level, modality, language). The framework yields explainable, auditable scores with traceable contributions by category and exemplary comments, supporting both accountability and improvement. By operationalizing student voice with methodological safeguards, the PULSE Index offers a rigorous alternative to traditional department rankings and a practical blueprint for institutions seeking evidence-based, student-centered quality assurance. An illustrative worked example demonstrates implementation details and reporting.