Deep Convolutional Neural Networks for Spatiotemporal Crime Prediction

Crime, as a long-term global problem, has been showing the complex interactions with space, time and environments. Extracting effective features to reveal such entangled relationships to predict where and when crimes will occur, is becoming a hot topic and also a bottleneck for researchers. We, therefore, proposed a novel Spatiotemporal Crime Network (STCN), in an attempt to apply deep Convolutional Neural Networks (CNNs) for automatically crime-referenced feature extraction. This model can forecast the crime risk of each region in the urban area for the next day from the retrospective volume of high-dimension data. We evaluated the STCN using felony and 311 datasets in New York City from 2010 to 2015. The results showed STCN achieved 88% and 92% on F1 and AUC respectively, confirming the performances of STCN exceeded those of four baselines. Finally, the predicted results was visualized to help people understanding its linking with the ground truth.

[1]  Bruce A. Draper,et al.  Automatically Searching for Optimal Parameter Settings Using a Genetic Algorithm , 2011, ICVS.

[2]  Tao Cheng,et al.  Novel evaluation metrics for sparse spatio-temporal point process hotspot predictions - a crime case study , 2016, Int. J. Geogr. Inf. Sci..

[3]  Gregory Shakhnarovich,et al.  FractalNet: Ultra-Deep Neural Networks without Residuals , 2016, ICLR.

[4]  Naftali Tishby,et al.  Opening the Black Box of Deep Neural Networks via Information , 2017, ArXiv.

[5]  Fernando Nogueira,et al.  Imbalanced-learn: A Python Toolbox to Tackle the Curse of Imbalanced Datasets in Machine Learning , 2016, J. Mach. Learn. Res..

[6]  Yu Zheng,et al.  Deep Spatio-Temporal Residual Networks for Citywide Crowd Flows Prediction , 2016, AAAI.

[7]  Xinyue Ye,et al.  Patterns of Near-Repeat Gun Assaults in Houston , 2012 .

[8]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[9]  Theodoros Damoulas,et al.  Mining 911 Calls in New York City: Temporal Patterns, Detection, and Forecasting , 2015, AAAI Workshop: AI for Cities.

[10]  Addarsh Chandrasekar,et al.  Crime Prediction and Classification in San Francisco City , 2015 .

[11]  Jane Law,et al.  Bayesian Spatio-Temporal Modeling for Analysing Local Patterns of Crime Over Time at the Small-Area Level , 2014 .

[12]  K. Leong,et al.  A review of spatio-temporal pattern analysis approaches on crime analysis , 2015 .

[13]  Mohsen Kalantari,et al.  Spatio-temporal analysis of crime by developing a method to detect critical distances for the Knox test , 2016, Int. J. Geogr. Inf. Sci..

[14]  Leslie W. Kennedy,et al.  Risk Terrain Modeling: Brokering Criminological Theory and GIS Methods for Crime Forecasting , 2011 .

[15]  Daniel Kifer,et al.  Crime Rate Inference with Big Data , 2016, KDD.

[16]  Matthew S. Gerber,et al.  Predicting crime using Twitter and kernel density estimation , 2014, Decis. Support Syst..

[17]  Irena Pletikosa Cvijikj,et al.  Exploring Foursquare-derived features for crime prediction in New York City , 2016 .

[18]  Peng Wang,et al.  Semantic Clustering and Convolutional Neural Network for Short Text Categorization , 2015, ACL.

[19]  Sang Michael Xie,et al.  Combining satellite imagery and machine learning to predict poverty , 2016, Science.

[20]  Xiaofeng Wang,et al.  Spatio-temporal modeling of criminal incidents using geographic, demographic, and twitter-derived information , 2012, 2012 IEEE International Conference on Intelligence and Security Informatics.

[21]  Yair Movshovitz-Attias,et al.  Ontological supervision for fine grained classification of Street View storefronts , 2015, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[22]  D Carlos.  Early Warning System for Temporary Crime Hot Spots , 2016 .

[23]  Yunpeng Wang,et al.  Large-Scale Transportation Network Congestion Evolution Prediction Using Deep Learning Theory , 2015, PloS one.

[24]  Xiang Li,et al.  Deep learning architecture for air quality predictions , 2016, Environmental Science and Pollution Research.