多重杀人犯罪动机测量模型的创建

The Establishment of Prediction Model of Criminal Motive of Multiple Murder

精准预测多重杀人犯罪动机对于案件侦破方向、范围的确定和有效预防战略的实施来说有着非同寻常的实践意义。为创建多重杀人犯罪动机测量模型,采用阶层抽样、整群抽样方法选取自1980至2016年间由单独个体实施的多重杀人案例181件。采用开放式编码、主轴编码和重点编码方法对所收集的案例文本逐一进行编码,得到4个主范畴名下13项自变量因素指标。基于对181件案例的经验分析和相关文献的归纳总结,提出多重杀人犯罪动机4种不同的分类,得到犯罪动机因素4个不同水平。将犯罪动机因素作为因变量,每个主范畴名下的因素指标作为自变量,分别创建4个多项logistics回归模型。根据模型的参数估计,可计算出任一多重杀人凶犯作案动机类型的概率值。经检验,模型对追求刺激和快乐、追求利益、心怀愤恨、权力和信念等4种犯罪动机类型的预测准确率分别为76.7%、75.0%、93.8%、38.5%。

It is very significant to accurately predict criminal motive of multiple murder for the direction and range of breaking a case. In order to create the prediction model of criminal motive of multiple murder,we selected 181 samples of the names of multiple murder cases by lonely individual from 1980-2016 by means of cluster sampling and stratified random sampling. Through text analysis we coded all the text step by step. By using open coding,main axis of coding and selective coding,thirteen independent variables under four main categories were selected. By means of analyzing 181 multiple murder cases empirically and thesis data comprehensively,four categories of motive of multiple murder were put forward at first. Therefore four levels of criminal motive of multiple murder were obtained. Regarding criminal motive of multiple murder as dependent variables and elements of every main categories as independent variables,four multinomial logistics regression models were created. Therefore criminal plot（ x1）,corpse dealt with（ x3）,aim acts（ x4）,The social status of victims（ x9）,whether the individual emotion is not placed safely for long（ x12） and whether there is a criminal record（ x13） and so on were selected. On the basis of these multinomial logistics regression model of motive of multiple murder was created. According to the model,probabilities of each category of criminal motive of multiple murder can be calculated. After testing,the prediction accuracy rate of the model is 76. 7% for pursuit of stimulation and happiness,75. 0% for pursuit of interests,93. 8% for resentment,38. 5% for power and belief.