ScholarWorks@숙명여자대학교

초록

Purpose This paper aims to develop a no-show prediction model based on behavioral reservation data and to design a risk-based differential pricing system to improve the operational efficiency of reservation-based services. By integrating prediction, customer segmentation, and policy simulation, this study proposes a risk-based pricing framework applicable to both public and private service settings. Methods Using a public dataset of 110,527 medical appointments from Brazil, a logistic regression model was used to estimate individual no-show probabilities. Key behavioral variables—including age, gender, SMS reminders, reservation–visit interval, and past no-show history—were analyzed. Predicted probabilities were then used to classify customers into risk groups and to simulate a risk-based differential pricing scheme. Results Age, SMS reminder receipt, and the reservation–visit interval were identified as significant predictors of no-shows. The model effectively classified customers into four risk levels with clear differences in observed no-show rates. Simulation results indicate that tailored pricing or prepayment policies for high-risk groups can reduce no-shows and improve service efficiency while maintaining fairness for low-risk customers. Conclusion This study demonstrates that behavioral data can effectively predict no-show risk and support the design of a risk-based differential pricing model. The proposed approach provides actionable insights for reducing no-shows, improving resource allocation, and enhancing fairness in reservation-based services.

키워드