Deep Learning for Real-Time Crime Forecasting and Its Ternarization
Deep Learning for Real-Time Crime Forecasting and Its Ternarization
复制标题
DOI：
10.1007/s11401-019-0168-y
复制
发表时间：
2017-11
期刊：
影响因子：
--
通讯作者：
Bao Wang;Penghang Yin;A. Bertozzi;P. Brantingham;S. Osher;J. Xin
中科院分区：
文献类型：
--
作者：
Bao Wang;Penghang Yin;A. Bertozzi;P. Brantingham;S. Osher;J. Xin
实时犯罪预测非常重要。然而，准确预测下一次犯罪将在何时何地发生是困难的。没有一个已知的物理模型能合理地近似描述这样一个复杂的系统。历史犯罪数据在空间和时间上都很稀少，利益信号很弱。在这项工作中，作者首先提出了一个适当的表示犯罪数据。然后，作者适应时空残差网络的良好代表的数据来预测犯罪的分布在洛杉矶在邻里大小的包裹的时间尺度。这些实验以及与现有的几种预测方法的比较表明，所提出的模型在准确性方面的优越性。最后，作者提出了一个三端化技术，以解决其部署在真实的世界的资源消耗问题。这项工作是我们的短会议论文[王，B.，Zhang，D.，中国农业科学院农业研究所所长，Zhang，L. H、例如，深度学习用于真实的时间犯罪预测，2017，arXiv：1707.03340]。
Real-time crime forecasting is important. However, accurate prediction of when and where the next crime will happen is difficult. No known physical model provides a reasonable approximation to such a complex system. Historical crime data are sparse in both space and time and the signal of interests is weak. In this work, the authors first present a proper representation of crime data. The authors then adapt the spatial temporal residual network on the well represented data to predict the distribution of crime in Los Angeles at the scale of hours in neighborhood-sized parcels. These experiments as well as comparisons with several existing approaches to prediction demonstrate the superiority of the proposed model in terms of accuracy. Finally, the authors present a ternarization technique to address the resource consumption issue for its deployment in real world. This work is an extension of our short conference proceeding paper [Wang, B., Zhang, D., Zhang, D. H., et al., Deep learning for real time Crime forecasting, 2017, arXiv: 1707.03340].