基于机器学习的恐怖袭击事件后果预测方法研究

恐怖袭击事件通常会造成严重的人员伤亡、财产损失和社会影响，针对在不同场景下发生恐怖袭击所造成的后果进行预测是目前应对恐怖袭击事件急需解决的问题之一。利用多源数据，首先基于随机森林算法对恐怖袭击事件是否造成死伤进行分类预测，进而基于岭回归算法预测事件造成的具体死伤人数。研究结果表明:随机森林在测试集上对有死伤事件的召回率达到0.85，岭回归预测死亡和受伤人数的平均绝对误差分别小于1人和2人。研究结果可为反恐资源配置优化、预防恐怖袭击事件和减少其造成的损害提供辅助决策支持。

Study on method of consequence prediction for terrorist attacks based on machine learning

The terrorist attacks usually cause serious casualties,property loss and social impact,so the prediction of consequence caused by the terrorist attacks under various scenarios is one of the problems which urgently need to be solved for the response of terrorist attacks at present.Through using the multi source data,the classified prediction on whether the terrorist attacks would cause casualties was carried out based on the random forest algorithm,and the specific casualties caused by the terrorist attacks were predicted by using the ridge regression algorithm.The results showed that the recall rate for the casualties events of random forest on the testing sets reached 0.85,and the mean absolute error of the ridge regression prediction on the death toll and the number of injury was less than 1 and 2 respectively.The results can provide decision making support for optimizing the allocation of counter terrorism resources,as well as preventing and reducing the terrorist attacks and the damage caused by them.