CAJ | 학술논문

为了对恐怖事件实现早期预警，通过分析恐怖事件的网络结构，构建了恐怖事件的预测模型。利用隐马尔可夫模型与贝叶斯网络方法，通过分析一些先前发生的事件来预测恐怖分子在未来一段时间可能发动的恐怖活动，实现对相关情报的侦测，预防可能发生的恐怖事件。同时，对完备数据与不完备数据条件下的恐怖事件的预测算法进行分析。结果表明，提出的预测方法与使用监测软件得到的结果相近，验证了隐马尔可夫模型的合理性和贝叶斯网络方法的有效性。不足之处在于，监测过程获取的情报信息较少，一定程度上影响了模型结果的精确度。
위료대공포사건실현조기예경，통과분석공포사건적망락결구，구건료공포사건적예측모형。이용은마이가부모형여패협사망락방법，통과분석일사선전발생적사건래예측공포분자재미래일단시간가능발동적공포활동，실현대상관정보적정측，예방가능발생적공포사건。동시，대완비수거여불완비수거조건하적공포사건적예측산법진행분석。결과표명，제출적예측방법여사용감측연건득도적결과상근，험증료은마이가부모형적합이성화패협사망락방법적유효성。불족지처재우，감측과정획취적정보신식교소，일정정도상영향료모형결과적정학도。
To finish early-warning for terrorism incident,the network structure of terrorism incident was an-alyzed,and predicting model constructed.Hidden Markov Model and Bayesian networks were proposed to construct terrorism incident predicted model.Some possible terrorism activities were tracked by analyzing some previous events for predicting other terrorism activist in the future.By predicting these activities,in-formation could be obtained to forecast ahead of schedule,and prevent some possible terrorism accidents. Terrorism incident prediction algorithms were analyzed under perfect data and imperfect data circumstance. The experimental results show that the prediction method and monitoring software of the actual monito-ring results are very similar and validate the rationality of hidden Markov models and Bayesian network method.The shortage is less information acquisition in the process of monitoring,which to a certain de-gree also affects the accuracy of the results of the model.