Daily tourist flow forecasting using SPCA and CNN-LSTM neural network

摘要

Predicting the daily tourism flow of scenic spots is of great significance for improving the management quality and the tourist experience. Affected by complex factors, daily tourism flow data have strong nonlinear characteristics. In this article, a multilayer neural network S-CNNLSTM is put forward to make accurate short-term tourism flow prediction. First, to reduce the redundant information between the influencing factors, sparse principal component analysis is adopted to reduce the data dimension. Then the processed data is input into a deep neural network framework that combines the convolutional neural network (CNN) and long short-term memory (LSTM) network. CNN extracts local trends, and LSTM is introduced to learn the inner law of time series and make prediction. Finally, through the experiments with real data and the comparison algorithms, the stability and practicability of the proposed method are verified.