Study On The Forecasting Of Passenger Checked Baggage Demand For Departure Flights

Aiming at the problems of insufficient utilization of passenger cabin capacity and low capacity,this paper puts forward a set of baggage demand estimation mechanism for passengers on departing flights to provide scientific data support for baggage loading personnel in order to ensure that airlines can improve the utilization rate of passenger cabin.Aiming at the problem of missing data,an optimization algorithm based on kNN-DBSCAN missing data filling is proposed.In order to make a better scientific prediction for the departure flight cabin baggage stowage and so on,this paper preprocesses the data information of baggage demand obtained through the "self-help baggage check system" equipment,including the deletion of duplicate data and the interpolation of missing data.Aiming at the interpolation problem of missing data,a kNN-DBSCAN algorithm for filling missing data is proposed.The density based DBSCAN clustering algorithm is applied to the kNN nearest neighbor filling algorithm,and the real data is used to verify the algorithm.The results show that the algorithm is better than the traditional kNN algorithm to fill the missing data.The characteristics of checked baggage demand of departing passengers are analyzed.Passenger baggage demand forecast for departing flights is affected by multiple factors and levels.This paper analyzes the conventional and unconventional factors that affect baggage demand prediction,selects the factors with strong correlation type as feature vector,and statistics the baggage demand data of different time periods.On this basis,the time distribution characteristics of checked baggage demand are deeply excavated,and the research shows that baggage demand has daily similarity.A long-term and short-term forecasting model of passenger baggage demand for departing flights is constructed.Based on the data of baggage demand history,a short-time forecasting model based on BP neural network is constructed to predict the baggage demand in a certain period of time.The experimental results show that the long-term and short-term forecasting models are scientific,with high prediction accuracy and strong robustness.By establishing the forecast model of passenger checked baggage demand for departing flights,scientific data support for baggage loaders is provided to ensure that airlines improve the utilization rate of existing passenger cabins.