Prediction of uncertain passenger flow in scenic spots by fusing multi-source data and integrated learning

Accurate scenic spot traffic prediction is of great significance for the optimal allocation of tourism resources and safety management.Aiming at the shortcomings of traditional methods in coping with data multi-source and prediction uncertainty, this study proposes an uncertainty prediction framework that integrates multi-source data and integrated learning.By integrating heterogeneous data from multiple sources, such as historical passenger flow, meteorology, web search and spatial features, a heterogeneous integrated model based on random forest, XGBoost and long short-term memory (LSTM) is constructed, and quantification of uncertainty is realised by combining quantile regression and conformal prediction method.Experiments on public datasets show that this method reduces the mean square error (MSE) by 30%, the mean absolute percentage error (mean absolute percentage error) by 25%, and the prediction interval coverage reaches 95.3%, which provides reliable decision support for the intelligent management of scenic spots.