马尔科夫链在对话信息识别犯罪中的应用  

The Application of Markov Chain in Identifying Crime by Dialogue Information

作　　者：张俊丽[1] 

机构地区：[1]西安欧亚学院通识教育学院,陕西西安710065

出　　处：《数学的实践与认识》2015年第22期176-181,共6页Mathematics in Practice and Theory

摘　　要：通过对话信息进行识别分析,构造了罪犯嫌疑度的一种确定方法.首先对人群的对话模式进行了分类,确定了9种对话类型下罪犯的嫌疑程度权向量.然后计算出话题权向量的初始值,再引入修正向量对话题权向量进行修正.最后利用马尔科夫法对人员犯罪的嫌疑度进行了分析.通过每个人传递敏感信息的信息量,构造概率转移矩阵,从而计算出在长期作用下将会出现的稳定状态,即稳定状态时人员犯罪的嫌疑度.This article classifies the message transfer nodes of three groups and determines the conspiracy weight vector of 9 kinds of mode.And after figuring out the initial topic weight vector,we also introduce modifying vector to update the initial topic weight.Markov Chain is applied in the steady situation which appears under long-term effects.By introducing the number of sensitive message that everyone transmitted,the article constructs the transition probability matrix.Finally,the conspiracy rates of staffs can be determined.

关 键 词：马尔科夫链 权向量 信息识别 概率转移矩阵 

分 类 号：O211.62[理学—概率论与数理统计]