Research On Multi-source Heterogeneous Data Fusion Method Of Civil Aviation Revenue System Based On Semantics

Air ticket agents make use of loopholes in civil aviation revenue to generate a large number of invalid bookings,causing great losses to airlines.At present,the civil aviation revenue system cannot completely avoid the losses caused by revenue loopholes.In order to fundamentally solve the problem of revenue loopholes,only relying on advanced data mining technology to find suspicious orders and plug the loopholes.The massive multi-source heterogeneous data in the civil aviation revenue system has the characteristics of multi-modality,high dimensions,and many missing values,which causes great difficulties to traditional data mining tasks.In order to efficiently complete data mining tasks,this paper focuses on multi-modal data fusion and uses advanced data fusion techniques to complete multi multi-modality data mining tasks.This paper first proposes a large-scale multi-modal data fusion algorithm(RMSKMC),which finds the optimal subspace on a single modal to achieve the self-reducing dimension of high-dimensional data,and uses non-negative matrix factorization(NMF)to reduce the loss Reconstruction is performed so that different modalities share the same clustering indicator matrix to achieve multimodal information complementation,complete large-scale multimodal data fusion,and finally combine K-means algorithm to complete data mining tasks.The experimental results show that the algorithm consumes less resources than other multimodal fusion algorithms on large-scale multimodal data sets,and has better fusion performance to finally achieve better clustering results.On this basis,due to the large number of missing values in the multi-source heterogeneous data set of the civil aviation revenue system,which greatly affects the data fusion performance,combined with the above fusion algorithm,a semantic-based multi-source heterogeneous data fusion algorithm for the civil aviation revenue system is further proposed.First,the traditional MCEM algorithm is improved to make it suitable for the civil aviation revenue system data set,greatly speeding up its convergence speed,and it can quickly complete a large number of missing data in the civil aviation revenue system data,and then use RMSKMC algorithm for fusion clustering.The experimental results show that the algorithm converges fast on the multi-modal data set of the civil aviation revenue system,and can better identify suspicious orders in the civil aviation revenue system.This article introduces multi-modal data fusion into the revenue loophole problem of civil aviation revenue system for the first time,using multi-source heterogeneous data fordata fusion,combined with clustering algorithm,can accurately predict suspicious orders,making it possible to reduce the loss caused by the revenue loophole to zero,Pointed out the direction for future related research.