A potential value mining method for flight departure time based on LDA model

Effective analysis of air route value will help the airport managers to evaluate their operational position. According to the calculation method of air route value, the indicator of flight departure time is selected in this manuscript to mine the potential value of flight departure time based on Latent Dirichlet Allocation (LDA) model. Firstly, obtaining selection of the air route from the passenger departure data through Gibbs Sampling method. Then, vectorizing the result of air route selection and simplifying the value of flight departure time. Finally, the LDA model of passenger-air route- departure time is established to mine the potential value of flight departure time. Experimental results on the passenger departure data of a large domestic airport in 2018 demonstrate that the LDA model can more efficiently obtain correlation of departure time than the Latent Semantic Indexing (LSI) model.