A Study On The Tourism Demand Forecasting Using Echo State Networks With Improved Fruit Fly Optimization Algorithm

With the improvement of national consumption level,the tourism industry plays an increasingly important role in the national economy,and the proportion of tourism revenue to GDP continues to increase.The increase in the number of tourists puts higher demands on the management of tourist cities and attractions.Tourism demand is affected by multiple factors such as seasons,holidays,weather and emergencies.How to accurately predict the number of passengers,for industry regulators and operators,is of great significance.Many scholars have tried to use different models to predict tourism demand.Artificial intelligence methods have great advantages over traditional methods in adaptive learning ability and nonlinear fitting ability.In recent years,they have become the focus of academic research.Firstly,this thesis improves the local search ability and search efficiency of the standard Fruit Fly Optimization Algorithm(FOA),by introducing adaptive factors for the population of the fruit fly and the search step size,and optimizing the initial iteration position.Through combining the optimized FOA and Echo State Networks(ESN),a two-stage assembled forecast model named AFOA-ESN is proposed.The AFOA optimizes the key parameters of ESN,and the optimized parameters are put into the ESN.Finally,monthly data of Beijing City and Hainan Province overnight passengers are selected to test the performance of AFOA-ESN.The experimental results show that AFOA-ESN model has higher prediction accuracy than autoregressive moving average model,support vector machine,BP neural network,standard ESN network and other prediction models.At the same time,the convergence speed and prediction accuracy of AFOA-ESN are better than standard ESN and FOA-ESN,which proves the effectiveness of the proposed AFOA-ESN model.