Research And Application Of Inferring The Travel Purposes Of Civil Aviation Passenger Groups

People usually travel together with others to a common destination and for a common purpose. Inferring the travel purposes of passenger groups is a very interesting research problem because it should bring about meaningful changes for passenger service and decision-making of passenger carriers, organizations and even governments. In this paper, we attempt to solve this problem by using the historical travel records of passengers. For overcoming the constraint of IID assumption of traditional classifiers, we propose a novel collective inference method based on the idea of iterative classification.First, we construct co-travel networks by extracting social relations between passengers from their travel records available in carriers’passenger information systems. Based on co-travel networks, we generate a series of features for passenger groups to capture the dependencies between their labels. These features are then used indicators preliminary pre-classification of passenger groups. After pre-classification, we use the overlap between the different passenger groups to form a class label based the idea of collective classification, then we can begin iteration classification with these label, the iteration classification would end until the results are no longer change.Experimental results on a real dataset of passenger travel records in the field of civil aviation demonstrate that the proposed iterative classification method can efficiently infer the travel purposes of passenger groups. At last, we infer the travel purpose of passenger groups from different city, then analysis the city character and development trend by the result of inferring.