Optimization of Feature Selection on Student Complaint Data Using Recursive Feature Elimination to Improve Academic Service Quality

Abstract

This study aims to optimize the feature selection process on student complaint data regarding academic services in universities using the Recursive Feature Elimination (RFE) method. Student complaints' diverse and complex nature requires in-depth analysis to identify crucial features that affect service satisfaction. An accurate feature selection process can help universities understand the most frequently reported issues, enabling them to respond and improve services more effectively. The research utilizes complaint data from various aspects of academic services, such as administration, facilities, and faculty interactions. After preprocessing the data to remove noise and irrelevant entries, RFE is applied to select the most relevant features. Subsequently, a classification model is built using the selected features to identify the most significant complaint patterns. Model evaluation is conducted through cross-validation techniques to ensure accuracy and reliability, with metrics such as accuracy, precision, recall, and F1-score. The results demonstrate that the RFE method significantly enhances model performance in selecting essential features, making the classification model more efficient and accurate in predicting student complaints. Thus, this study contributes significantly to assisting universities in enhancing the quality of academic services through a more targeted analysis of student complaints. Implementing this method will improve the complaint-handling process and increase overall student satisfaction