Machine Learning for Emergency Service Optimization: A Real-World Application

The paper introduces an advanced decision support system based on machine learning to optimize emergency service allocation in hospital emergency departments (EDs). By predicting unscheduled patient arrivals and the required number of physicians based on real-world data, our system addresses the complexity of ED operations, influenced by internal factors such as resource availability and external environmental unpredictability. We address the problem by enabling hourly and daily forecasting for emergency departments, incorporating the lagged effects of environmental pollution shifted by months. Comparative evaluations are conducted using statistical and machine learning methods to verify the performance of our approach in minimizing patient waiting times and costs for emerging cases. This study demonstrates the effectiveness of machine learning in enhancing decision-making processes and optimizing emergency services within dynamic healthcare environments.