Risk Prediction Of Theft Crimes Using Lstm And St Gcn
Keywords:
LSTM ,GCN , ST-GCN, RMSE , MAPE , R2_Score , Crime rate prediction , risk of increasing crimes.Abstract
Growing population increasing urbanization which
will give rise to growing communities and will
increase risk of crimes. Increasing crime will put
people’s security at risk and make difficult for
security professional to maintain law and order. To
tackle crime in growing urban communities’ author
of this paper employing combination of two different
algorithms such as LSTM (long short-term memory)
and GCN (Graph Convolution Network). LSTM is
known to perform better prediction on temporal
features like Date & Time and GCN known for best
prediction on Spatial Features like Location.
Combination of both LSTM & GCN can
automatically and effectively detect the high-risk
areas in a city. Topological maps of urban
communities carry the dataset in the model, which
mainly includes two modules such as spatial &
temporal features extraction module and temporal
feature extraction module to extract the factors of
theft crimes collectively. To train and test above
algorithm performance author has utilized US crime
and Chicago Crime datasets and then for spatial
features author has joined weather information like
Temperature on Chicago dataset but this weather
information extraction process we don’t know so we
have used Date & Time for LSTM temporal features
training and Crime locations as Spatial features
training with GCN. Propose algorithm performance
in terms of R2square, RMSE and MAPE. R2square
consider as accuracy for crime rate prediction
model and then RMSE (root mean square error) and
MAPE (mean absolute percentage error) represents
difference between true and predicted crime rate. So
the lower the difference the better is the algorithm.
Combination of propose LSTM & GCN algorithm is
known as ST-GCN and compare this propose
algorithms performance with existing algorithms
like
Random Forest and LSTM. Among all
algorithms propose is giving high R2score and less
RMSE and MAPE.
Downloads
References
[1] U. Habitat, State of the Worlds Cities
2012/2013: Prosperity of Cities. Evanston, IL,
USA: Routledge, 2013.
[2] F.Cicirelli, A. Guerrieri, G. Spezzano, and
A. Vinci, An edge-based platform for dynamic
smart city applications, Future Gener. Computer.
System., vol. 76, pp. 106118, Nov. 2017.
[3] M. A. Tayebi, M. Ester, U. Glässer, and P. L.
Brantingham, CRIME TRACER: Activity spaced
based crime location prediction, in Proc.
IEEE/ACM Int. Conf. Adv. Social Netw. Anal.
Mining (ASONAM), Aug. 2014, pp. 472480.
[4] H. Wang, D. Kifer, C. Graif, and Z. Li, Crime
rate inference with big data, in Proc. 22nd ACM
SIGKDD Int. Conf. Knowl. Discovery Data
Mining, Aug. 2016, pp. 635644.
[5] A. Crawford and K. Evans, Crime prevention
and community safety, Oxford Univ. Press,
Oxford, U.K., Tech. Rep. 35, 2017.
[6] M. Friedman, A. C. Grawert, and J. Cullen,
Crime trends, 1990 2016, Brennan Center Justice,
New York Univ. School Law, New York,
NY,USA, Tech. Rep., 2017.
[7] X. Hu, J. Wu, P. Chen, T. Sun, and D. Li,
Impact of climate variability and change on crime
rates in Tangshan, China, Sci. Total Environ., vol.
609, pp. 10411048, Dec. 2017
[8] M. Kassen, A promising phenomenon of
open data: A case study of the chicago open data
project, Government Inf. Quart., vol. 30, no. 4, pp.
508513, Oct. 2013.
[9] P. Homel and N. Masson, Partnerships for
urban safety in fragile contexts: The intersection
of community crime prevention and security
sector reform, Geneva Peacebuilding Platform,
Geneva, Switzerland, White Paper 17, 2016.
[10] S.-C. Shi, P. Chen, P.-H. Yuan, C. Hou, and
H.-X. Ming, The prediction of offender identity
using decision-making tree algorithm, in Proc. Int.
Conf. Mach. Learn. Cyber. (ICMLC), vol. 2, Jul.
2018, pp. 405 409.
[11] M. Lim, A. Abdullah, N. Jhanjhi, and M. K.
Khan, Situation-aware deep reinforcement
learning link prediction model for evolving
criminal networks, IEEE Access, vol. 8, pp.
1655016559, 2020.
[12] R Kumar and B. Nagpal, Analysis and
prediction of crime patterns using big data, Int.
J. Inf. Technol., vol. 11, no. 4, pp. 799805, Dec.
2019.
[13] P. Das, A. K. Das, J. Nayak, D. Pelusi, and
W. Ding, A graph based clustering approach for
relation extraction from crime data, IEEE Access,
vol. 7, pp. 101269101282, 2019.
[14] Q. Zhang, P. Yuan, Q. Zhou, and Z. Yang,
Mixed spatial-temporal char acteristics based
crime hot spots prediction, in Proc. IEEE 20th Int.
Conf. Computer. Supported Cooperate. Work
Design (CSCWD), May 2016, pp. 97101.
[15] Y. Mei and F. Li, Predictability comparison
of three kinds of robbery crime events using
LSTM, in Proc. 2nd Int. Conf. Data Storage Data
Eng. (DSDE), 2019, pp. 2226.
[16] B. J. Jefferson, Predictable policing:
Predictive crime mapping and geographies of
policing
and race, Ann. Amer. Assoc.
Geographers, vol. 108, no. 1, pp. 116, Jan. 2018.
[17] B. Wang, P. Yin, A. L. Bertozzi, P. J.
Brantingham, S. J. Osher, and J. Xin, Deep
learning for real-time crime forecasting and its
ternarization, Chin. Ann. Math. B, vol. 40, no. 6,
pp. 949966, Nov. 2019.
[18] A. Ghazvini, S. N. H. S. Abdullah, M. K.
Hasan, and D. Z. A. B. Kasim, Crime
spatiotemporal prediction with fused objective
function in time delay neural network, IEEE
Access, vol. 8, pp. 115167115183, 2020.
[19] S. D. Johnson, L. Summers, and K. Pease,
Offender as forager? A direct test of the boost
account of victimization, J. Quant. Criminal., vol.
25, no. 2, pp. 181200, Jun. 2009.
[20] L.E. Cohenand M. Felson, Social change and
crime rate trends: A routine activity approach,
Amer. Social. Rev., vol. 44, no. 4, pp. 588608,
1979.
[21] R. V. G. Clarke and M. Felson, Routine
Activity and Rational Choice. Piscataway, NJ,
USA: Transaction Publishers, 1993.
[22] N. B. Davies, J. R. Krebs, and S. A. West, An
Introduction to Behavioural Ecology. Hoboken,
NJ, USA: Wiley, 2012.
[23] D. K. Rossmo, Geographic Pro ling. Boca
Raton, FL, USA: CRC Press, 1999.
[24] S. D. Johnson and K. J. Bowers, The burglary
as
clue to the future: The beginnings of
prospective hot-spotting, Eur. J. Criminal., vol. 1,
no. 2, pp. 237255, Apr. 2004.
[25] C. P. Haberman and J. H. Ratcliffe, The
predictive policing challenges of near repeat
armed street robberies, Policing A, J. Policy
Practice., vol. 6, no. 2, pp. 151166, Jun. 2012.
[26] S. Block and S. Fujita, Patterns of near repeat
temporary and permanent motor vehicle thefts,
Crime Prevention Community Saf., vol. 15, no. 2,
pp. 151167, May 2013.
[27] W. Wells, L. Wu, and X. Ye, Patterns of near
repeat gun assaults in hous ton, J. Res. Crime
Delinquency, vol. 49, no. 2, pp. 186212, May
2012.
[28] J. Sturup, M. Gerell, and A. Rostami,
Explosive violence: A near-repeat study of hand
grenade detonations and shootings in urban
Sweden, Eur. J. Criminal., vol. 25, no. 4, 2019, Art.
no. 1477370818820656.
[29] J. Zhang, Y. Zheng, and D. Qi, Deep spatio
temporal residual networks for citywide crowd
ows prediction, in Proc. 31st AAAI Conf.
Artificatal. Intell., 2017, pp. 16551661.
[30] T. N. Kipf and M. Welling, Semi-supervised
classification with graph convolutional networks,
2016, arXiv:1609.02907. [Online]. Available:
http://arxiv.org/abs/1609.02907
[31] S. Yan, Y. Xiong, and D. Lin, Spatial
temporal graph convolutional networks for
skeleton-based action recognition, in Proc. 32nd
AAAI Conf. Artif. Intell., 2018, pp. 74447452.
