A Comprehensive Analysis Of Student Feedback On The Curriculum, Lecturers, And Facilities At STMIK Mardira Indonesia Using A Statistical Approach And Machine Learning Natural Language Processing

Abstract

Enhancing the quality of higher education necessitates a thorough review of student feedback. Nonetheless, STMIK Mardira Indonesia continues to handle input manually, which challenges its ability to manage substantial amounts of both quantitative and qualitative data effectively. As a result, the institution fails to draw on numerous useful insights from student feedback.

This research aims to develop an integrated system that autonomously analyzes student input using statistical and machine-learning methods. The technology displays outcomes via an interactive dashboard to facilitate data-driven decision-making for management. The research employs a hybrid technique, integrating CRISP-DM for data analysis and Machine Learning with the Prototype method for system development.

The data analysis encompasses descriptive statistics for quantitative data (Likert scale) and Natural Language Processing (NLP) employing a Linear SVM model for sentiment classification of qualitative data (comments). The research yields a web-based system prototype that effectively consolidates the collection and analysis of student input on the curriculum, instructors, and facilities. The Linear SVM sentiment classification model has outstanding performance, with an accuracy of 95.2% on the test dataset. Moreover, the interactive dashboard effectively illustrates these outcomes through dynamic filtering. Consequently, the integration of statistical and Machine Learning methodologies successfully converts raw feedback into organized insights, equipping STMIK Mardira Indonesia with a reliable instrument to sustainably improve educational service quality.