A Combining Model for Crowd Flow Prediction in Partition Coverage Wireless Network Scenarios

With the development of modern higher education, university campuses serve multiple functions for both faculty and students, including teaching, research, and daily life. These campuses commonly face issues such as high population density, frequent mobility, and complex composition. Accurate prediction of campus traffic environments is fundamental for scientific and efficient management. Predictive data can support various aspects of campus operations, including energy conservation, resource allocation, event scheduling, hazard prevention, and emergency response, making it of significant importance. This paper leverages the advantages of wireless network seamless roaming technology to obtain the number of people in different areas from wireless controllers. Based on this, we propose a crowd flow prediction model using ARIMA-CNN-LSTM. The model has been tested, and its accuracy meets expectations and surpasses traditional algorithms.