Research On Terrorist Attack Analysis Method Based On Probabilistic Graph Sampling Aggregation Network

Terrorist attacks have always been an important security threat to human survival and social stability.At present,the intensification of international conflicts leads to the escalation of terrorist attacks in terms of number and scale,and various types of terrorist attacks are emerging in an endless stream.However,the existing researches on the classification and early warning of terrorist attacks lack the supports of deep learning models,resulting in poor analysis results.In the era of rapid development of big data technology and deep learning methods,facing the increasingly severe terrorist attacks,it is urgent to use more intelligent methods to carry out analyses and researches on terrorist attacks.For the analyses and researches on terrorist attacks,this paper proposed a method of terrorist attacks analysis based on probabilistic graph sampling aggregation network.This method consists of two parts: the first part is the construction of probabilistic graph sampling aggregation network model,and the second part is the analysis process of terrorist attacks based on this model.Through the combination of the two parts to carry out the classification of terrorist attacks and early warning researches.In the probabilistic graph sampling aggregation network model,the model can calculate the neighbor values of different nodes in the network by a given method,and generate a list of neighbor values.Based on the list,the model considers the importance and diversity of the nodes.According to the density,the model can realize a sampling to generate the neighborhood of different nodes.After determining the neighborhood of the nodes,the model uses aggregation function to transform and aggregate the characteristic information of the neighborhood nodes.After multi-layer feature information extraction and aggregation,the model finally forms the embedded representation of nodes in the network.In the analysis process of terrorist attacks based on probabilistic graph sampling aggregation network model,it mainly realizes the classification and early warning of terrorist attacks.In the process of classification of terrorist attacks,according to the embedded representation of nodes generated by the model,we trained a complex classifier to complete the task of classification of terrorist attacks.In the process of early warning analyses of terrorist attacks,we abstracted each event as a node in the network,and calculated the similarity,correlation,and connection probability between event nodes respectively.Finally,we used a neural network to construct a non-linear classification function to realize the comprehensive analysis of the above three indicators,which used to study the problem of link prediction and complete the early warning analyses of terrorist attacks.Compared with the original model and other network representation learning models,the probabilistic graph sampling aggregation network model has a better performance.In this paper,the method of terrorist attack analysis based on probabilistic graph sampling aggregation network can better complete the tasks of terrorist attacks analysis.The experimental results show that the proposed method can make full use of the characteristics of terrorist attacks,dynamically adjust the range of aggregation neighborhood,and combine multiple indicators for judgment.Therefore,it can realize the efficient analysis of terrorist attacks and provide continuous supports for the antiterrorism intelligence system.