Potential High-Value Passengers Discovery by Random Walk on Passenger-Route Heterogeneous Network

For the situation widespread in the civil aviation industry that large numbers of potential high-value passengers was ignored or can not be analysis because of technical reasons by airlines, a potential high-value passengers discovery by random walk on passenger-route heterogeneous network method was been proposed. Simulates the behavior of airline passengers choose routes by building a heterogeneous network of passengers and routes, and random walk between the departure airport and destination airport to discover the potential flight demand of each passenger, and thus evaluate potential value of passengers more effectively. Experiment on the PNR data set of civil aviation reservation systems of China shows: the algorithm we proposed is not only effective for frequent passengers, but also for the common passengers, it can found the passengers that travel by plane infrequently at present, but have great demand and become frequent passengers in the future.