Recognition Algorithm of Transaction Risk Events in Automobile Financial Market Based on Deep Learning

With the rapid development of economy, people’s material living standard has obviously improved, and people’s demand for material has also become greater. At the same time, a series of traffic environmental problems and technical problems of traffic management have gradually become prominent. In many aspects, the usual traffic management methods are not enough to meet these outstanding problems. In order to solve these problems, the design of intelligent transportation system becomes more and more important. In this paper, the vehicle recognition method based on deep learning is studied. The depth residual network is selected as the basic classification network and improved, and its second convolution block is introduced into dense connection to strengthen the feature flow of different depth convolution layers in the network, so as to improve the feature expression ability of the network. At the same time, random weight averaging and label smoothing are introduced into the residual network. On the one hand, the convergence of the network is stable and the oscillation amplitude is smaller, and the problem of weight oscillation is solved. On the other hand, the generalization ability of the model is strengthened to ensure the accuracy of fine-grained classification. In this paper, through the method of transfer learning, using the knowledge learned from a large number of data in ImageNet database to optimize network parameters, the Stanford CMS Kramp-Karrenbauer 196 data set is marked with fine granularity, divided into training set and test set for model training and learning, and finally effective features are extracted for vehicle classification. The accuracy of the two improved networks can reach 87.4% respectively. The experimental results show that the accuracy of the proposed method is high, which verifies the effectiveness of the method.