Research on Anti-terrorism Intelligence Mining Method Based on Attention Neural Networks

In order to improve the intelligence and prediction accuracy of anti-terrorism intelligence mining, this paper applies a medium of attraction neural network based on attention mechanism and investors will own a Complete set of open Source Intelligence Data Mining and takes the analysis of casualties and terrorist organizations as examples and a prediction model of casualties and potential about organizations is constructed. It has had been experimentally verified that its prediction ability is significantly improved compared to support vector machine algorithm, logistic regression algorithm, multi-layer perceptron and attention neural network.