IMPACT OF PREDICTIVE MACHINE LEARNING MODELS ON OPERATIONAL EFFICIENCY AND CONSUMER SATISFACTION IN UNIVERSITY DINING SERVICES

Abstract

The integration of predictive machine learning (ML) models into university dining services offers a transformative opportunity to enhance operational efficiency and elevate consumer satisfaction through data-driven strategies. This study quantitatively evaluates the impact of ML-based predictive analytics on key performance metrics within campus dining operations. A comprehensive dataset encompassing order volume, food waste, staffing patterns, and satisfaction survey scores was collected across four university dining facilities before and after the implementation of predictive ML systems. Statistical analyses, including paired t-tests and regression modeling, revealed significant improvements in several operational dimensions. Specifically, food waste was reduced by an average of 27%, service wait times decreased by 19%, and inventory turnover efficiency improved by 22% post-implementation. Additionally, consumer satisfaction scores—measured through standardized Likert-scale instruments—increased significantly in categories related to meal availability, freshness, and perceived responsiveness to preferences. The predictive models utilized supervised learning algorithms, primarily decision trees and random forests, to anticipate demand patterns and optimize procurement and scheduling decisions. Results demonstrate the practical viability of ML applications in enhancing efficiency and aligning service delivery with consumer expectations in real-time. This research provides empirical evidence supporting the adoption of predictive ML in university food service management and offers a scalable framework for institutions seeking to modernize operations through data analytics.