Page 112 - Proceedings of the 2017 ITU Kaleidoscope
P. 112
2017 ITU Kaleidoscope Academic Conference
(ii) improve the quality and efficiency of the services ren- [4] International Association of Crime Analysts (IACA),
dered by government entities and decision makers. CriClust Crime Pattern Definitions for Tactical Analysis, Stan-
focuses on point (ii) above [15], which has a direct posi- dards, Methods, and Technology (SMT) Committee
tive impact on point (i). If crime pattern is not delivered White Paper, 2011.
timeously then crime control is hampered, coupled with little
[5] S. Lin and D. E. Brown, “An outlier based data associ-
or no capital outlay to acquire armed weapons and related
ation method for linking criminal incidents,” Decision
materials aggravates the challenge of crime. While we have
Support Systems, vol. 41, no. 3, pp. 604–615, March
not presented smart data analysis or the CriClust system as
2006.
a panacea, the solution presented in this research is more
than a case study and is applicable to other crime domains.
[6] E. R. Groff and N. G. La Vigne, “Forecasting the fu-
This can help in pro-actively improving public safety, partic-
ture of predictive crime mapping,” Crime Prevention
ularly in resource-constrained settings such as in developing
Studies, vol. 13, pp. 29–57, 2002.
nations.
[7] J. Ferreira J, P. Jo´ ao, and J. Martins, “GIS for crime
analysis: Geography for predictive models,” In Elec-
6. CONCLUSION AND FUTURE WORK tronic Journal of Information Systems Evaluation, vol.
15, no. 1, pp. 36–49, 2012.
The motivation for this research is the incessant challenge
[8] O. Isafiade and A. Bagula, “Citisafe: Adaptive spatial
to tackle crime faced by public safety agencies, particularly
pattern knowledge using Fp-growth algorithm for crime
in resource constrained settings such as in developing na-
situation recognition,” in Proceedings of the IEEE In-
tions, which is an impediment to realising smart city devel-
ternational Conference on Ubiquitous Intelligence and
opment targets. This research has successfully demonstrated
Computing. December 2013, pp. 551–556, IEEE.
that the appropriate use of a cost-effective user-centred soft-
ware solution (e.g CriClust) could significantly assist crime
[9] O. E. Isafiade and A. B. Bagula, Data Mining Trends
reduction in resource constrained settings. CriClust can as-
and Applications in Criminal Science and Investiga-
sist analysts in suspect prioritisation, predicting and respond-
tions, IGI global USA, 2016.
ing to patterns that anticipate crime before it happens. This
will consequently help to tackle under-performance in cer- [10] T. Wang, C. Rudin, D. Wagner, and R. Sevieri, “Finding
tain core responsibilities of the police and help to develop patterns with a rotten core: Data mining for crime series
evidence-based policies. with core sets,” Big Data, vol. 3, no. 1, pp. 3–21, March
As future research, the CriClust web-based knowledge sup- 2015.
port system could consider combining mining of text and vi-
[11] A. Borg, M. Boldt, N. Lavesson, U. Melander, and
sual information, following a more extensive consideration
V. Boeva, “Detecting serial residential burglaries us-
for promoting effective investigative solution. For instance,
ing clustering,” In Expert Syst. Appl, vol. 41, no. 11,
attributes relating to suspect information (e.g tattoo, masked)
pp. 5252–5266, 2014.
could perhaps be translated into visual information (identik-
its) to mine suspect information and gain better insight into [12] L. W. Evett, G. Jackson, D. V. Lindley, and D. Meuwly,
the crime data. Further improvements are in form of incorpo- “Logical evaluation of evidence when a person is sus-
rating the use of crowd-sourcing, mobile phones and Wire- pected of committing two separate offences,” Journal
less Sensor Networks (WSNs) to improve the automation of of Science and Justice, vol. 46, no. 1, pp. 25–31, Else-
crime data collection and analysis. vier 2006.
[13] F. H´ uffner, C. Komusiewicz, and M. Sorge, “Find-
REFERENCES ing highly connected subgraphs,” in Proceedings of
the 41st International Conference on Current Trends
[1] S. Harrendorf, M. Heiskanen, and S. Malby, “Interna- in Theory and Practice of Computer Science, Pec pod
tional statistics on crime and justice,” Journal, vol. 64, Snˇ eˇ zkou, Czech republic, January 2015, pp. 1–20.
pp. 1–176, Helsinki 2010.
[14] D. R. Karger and C. Stein, “A new approach to the
minimum cut problem.,” Journal of ACM, vol. 43, no.
[2] South African Police Service, “Together squeezing
4, pp. 601–640, 1996.
crime to zero,” Strategic Plan, vol. 2012, no. 14, pp.
1–58, 2014. [15] A. Milgram, “Why smart statistics are the key to
fighting crime,” 2013, TED Talk, Retrieved from:
[3] C. Hafedh, G. Ramon, A. Theresa, N. Taewoo, M. Sehl,
https://www.ted.com/talks/anne-milgram-why-smart-
J. Hans, W. Shawn, and N. Karine, “Understanding
statistics-are-the-key-to-fighting-crime (accessed on 5
smart cities: An integrative framework,” in Proceed- November, 2015).
ings of the Hawaii International conference on System
Sciences (HICSS), 2012, pp. 2289–2297.
– 96 –