Research on Canteen Pedestrian Flow Prediction Model Based on Improved Gradient Boosting Decision Tree

With the increasing demand for catering services in universities, accurately predicting the flow of people in canteens has become the key to improving service quality and management efficiency. This study constructs a canteen pedestrian flow prediction model based on an improved gradient boosting decision tree algorithm. Firstly, multidimensional features such as historical pedestrian flow data, weather and precipitation data, weekday information, and holiday information were collected and organized in the canteen. Through data preprocessing and feature selection, significant factors affecting pedestrian flow prediction were identified. Subsequently, gradient boosting decision trees were used for model training and optimization, and compared with statistical mining analysis algorithms, XGBoost, and LightGBM algorithms. The experimental results show that the proposed model has significant advantages in prediction accuracy and robustness, and can effectively predict the flow of canteen consumers, providing a scientific basis for the operation and management of university canteens.