Design And Implementation Of Low-Cost Carrier Passenger Demand Forecasting System Based On Deep Learning

The emergence and rapid development of low-cost carrier(LCC)have promoted regional contacts and exchanges.As the key point of revenue management of lowcost Carrier,passenger demand forecasting mainly depends on the experience of decision-makers at present,but cause such problems as difficult to quantify the forecasting results,low accuracy and untraceable forecasting process.In addition,the choices of LCC passengers are more diverse,and the differences between different passengers will have an impact on prediction of LCC passenger transport demand.Therefore,in this paper,use deep learning method combining with the portraits of LCC passenger groups to study the demand forecast of LCC passenger transport demand and propose the forecasting model of LCC passenger transport demand based on deep learning.Also,this paper design and implement forecasting system to predict LCC passenger transport demand based on the forecasting model.This paper analyzes the system requirements and determines the main function points of the system.The overall design of the system is carried out using three-layer architecture which is data storage,classification prediction and business system.Also design database architecture and four functional modules which is passenger portrait,passenger demand prediction,authority control and display module.Among them,design and implementation of passenger portrait and passenger transport demand prediction function module are based on the LCC passenger transport demand prediction model proposed in this paper.The prediction model is analyzed and designed.The model mainly includes two parts: LCC passenger portrait and passenger demand prediction model.In the LCC passenger portrait part,PSO-K mean clustering algorithm is used to train the passenger portrait model and classify the LCC passenger group.In the passenger transport demand prediction model part,GRU neural network model integrating attention mechanism is built based on the passenger portrait results,and the model is Trained through Tensorflow to quantitatively predict the passenger transport demand change process of different passenger groups.Use Spring Boot and Vue framework to build the system.Realize the background logic processing,interface data interaction,visual display of LCC passenger transport demand forecast system.Through test,testify the system functions whether meet the design requirements and achieve the desired results.The completion of the LCC passenger transport demand forecasting system has fundamentally changed the current situation that the industry generally relies on the experience of practitioners to forecast passenger transport demand,and has realized the automation and systematization of the forecasting process.Therefore,providing important support for low cost airlines to optimize revenue management.