Queue time predictor: An AI-driven hybrid time series forecasting system for intelligent queue management

Estimating customer wait times is a constant challenge in service industries like banking, healthcare, retail, and government. Most queue management tools used today only show the current queue status. They do not provide any predictions, which leaves customers unsure about their wait times and prevents service providers from effectively allocating resources. This paper introduces a Queue Time Predictor that combines historical pattern analysis using the Auto Regressive Integrated Moving Average (ARIMA) model with real-time queue monitoring through a weighted hybrid forecasting strategy. The entire system was developed as a full-stack web platform using React.js, Node.js, MongoDB, and a Python-based forecasting microservice. Experimental evaluation showed that the hybrid model achieved a Mean Absolute Error of 3.18 minutes, compared to 5.42 minutes for the standalone ARIMA forecasting. This represents a 41.3% improvement in prediction accuracy. The system performed consistently well across different traffic scenarios, including steady flow, gradual build-up, sudden demand spikes, and mixed patterns.