论文部分内容阅读
提出了一种基于时空影响范围的网络构造方法,构造了一种基于节点影响强度的犯罪传输网络,并引入了复杂网络的度、平均度、聚集系数等特征参数分析犯罪传输网络。提取了犯罪预测过程中需要关注的重要节点,分析了其时间分布和空间分布特性,研究结果表明:1近邻的时空单元的犯罪率具有一定的关联关系。其中,节点的出度与入度具有正相关性,因此可以引入邻居时空单元的犯罪密度以量化和分析犯罪规律。2节点的度分布具有无标度特性,犯罪较少的小区也可能出现度较大的节点,而节点的度与未来犯罪率具有较大的关联性。因此,即便犯罪率较低的小区也要关注节点的度变化情况。3犯罪聚集系数大小与未来犯罪率的变化具有一定的关联性,较高的聚集系数意味着未来犯罪状态的变化。
A network construction method based on spatiotemporal impact scope is proposed. A crime transmission network based on node influence strength is constructed, and the characteristics of complex networks such as degree, average degree and aggregation coefficient are introduced to analyze the criminal transmission network. The important nodes that need attention in the process of crime prediction are extracted, and their time distribution and spatial distribution characteristics are analyzed. The results show that: (1) The crime rate of the nearest neighbor space-time units has a certain relationship. Among them, the out-degree of nodes has a positive correlation with the entry degree, so the crime density of neighbor space-time units can be introduced to quantify and analyze the criminal law. The degree distribution of 2 nodes has scale-free characteristics, and cells with fewer crimes may appear nodes with greater degrees, while the degrees of nodes have a greater correlation with future crime rates. Therefore, even if the crime rate is low in the community should also pay attention to the node degree of change. There is a certain correlation between the size of crime aggregation coefficient and the change of crime rate in the future. The higher aggregation coefficient means the change of crime state in the future.