Page 1
The Relationship between Crime and CCTV
Installation Status by Using Artificial Neural Networks
Ahyoung Jung1, Changjae Kim2,
Dept. S/W Engr. Soongsil University
[email protected] , [email protected]
Abstract. In this study, correlations are found between crime and CCTV using
multiple regression analysis and Artificial Neural Network. Determine
alternative ways to reduce crime by identifying the number of CCTV
installations for strong crime committed by local regions. Through a Multiple
regression analysis, we suggested a model of a CCTV installation that
determines the relationship between powerful crime and CCTV, which can
effectively prevent violent crime and prevent the possibility of effective
prevention of the crime.
Keywords: CCTV, Artificial Neural Network, Multiple Linear Regression
1 Introduction
Social unrest is rising as the nation's violent crimes soar. The CCTVs are being used
to deter criminal crimes and identify crimes against criminals worldwide. The CCTV
is effective in preventing crime prevention, and is needed to prevent effective crime
prevention measures, considering the possibility of crime zones and crime prone
areas.
A theoretical basis for preventing crime prevention by installing CCTV cameras is
the prevention of crime prevention. The crime prevention techniques against crime
prevention are criminal crime prevention techniques that enable criminals to deter
criminal crimes by preventing criminal crimes and control of criminal crimes
committed by criminals in the mid-1990s. Moreover, preventive crime prevention
theory is not a social system improvement, but a preventive approach that relies solely
on reducing crime opportunities. Thus, the theory of crime prevention differs from the
crime of criminal criminology, which focuses on crime in the context of immediate
environmental circumstances, which are expected to focus on the immediate
environment, circumstances, and characteristics[1]. Circumstance crime prevention is
based on rational choice theory, criminal opportunity theory, and crime prevention
theories through environmental design[2].
The CCTVs are rapidly increasing CCTVs in the wake of the recent crime
prevention, and CCTV monitors, which have been investigating the Ministry of
Public Administration and Home Affairs in May 2015, are 12,5608 CCTV cameras,
and 72 percent of them are CCTV cameras. Also, the crime prevention effect is
Advanced Science and Technology Letters Vol.139 (FGCN 2016), pp.150-157
http://dx.doi.org/10.14257/astl.2016.139.34
ISSN: 2287-1233 ASTL Copyright © 2016 SERSC
Page 2
expected to increase further as the expansion of CCTV tapes for crime prevention and
the development of intelligent CCTV technology progresses rapidly at a faster
pace[3].
There are limitations to preventing crime and responding to crime, such as crime
zones in crime zones and crime prone areas.
In this study, we propose to suggest a CCTV installation model that identifies the
status of crime in areas where local crime is organized and effectively prevent crime
prevention. The composition of this research is as follows. We will explore the
existing literature related to this study and explore the regression model and the
regression model of the research in this study. Chapter 3 describes the variables and
models used in the models of this study. After analyzing and verifying the processes
in Chapter 4, we will conclude the conclusion in Chapter 5, Present and Future, and
finalize this thesis.
2 Related Study
2.1 Crime Prevention and CCTV
With the recent surge in violent crime, the number of crimes in Korea is increasing,
causing the nation's stability to rise to a record low of 18.3 percent, according to data
compiled by the National Statistical Office. The installation of CCTVs in crime
prevention centers has proven to be a crime prevention effect, and a total of 23,000
criminals are found each year at the 79 CCTV service centers national wide[4].
In the United States, CCTV is being installed to prevent crime prevention in urban
areas and residential areas, and more than three times more CCTVs have been
installed since 9/11. As the U.S.'s main goal of the U.S.-led national policy toward the
United States since 9.11, the use of CCTV as a tool for prevention of crime and the
use of CCTV as a tool for the prevention of crime has been expanded. In the UK,
CCTV cameras were first introduced in the mid 1980s to prevent the country from
becoming the world`s first movable soccer field, the first in the world since the
introduction of CCTVs in the world. Currently, the nation is installing CCTVs in most
of the EU countries, but it is currently being installed as a focal point for preventing
the establishment of national security and prevention of terrorism[5].
2.2 Multiple Linear Regression Analysis
Multiple regression analysis is a form of regression analysis implied by a statistical
analysis of a causal relationship between variables. The regression analysis describes
the relationship between the independent variable and the dependent variables that
contribute to the resulting cause. Linear regression analysis is a linear regression
analysis for a linear regression model and a linear regression model with two
independent variables. Estimates of regression analysis provide estimates by
extrapolation of estimated regression models given the estimated regression model.
Advanced Science and Technology Letters Vol.139 (FGCN 2016)
Copyright © 2016 SERSC 151
Page 3
Linear regression models are often fitted using the least squares approach, but they
may also be fitted in other ways, such as by minimizing the "lack of fit" in some other
norm (as with least absolute deviations regression), or by minimizing a penalized
version of the least squares loss function as in ridge regression (L2-norm penalty) and
lasso (L1-norm penalty). Conversely, the least squares approach can be used to fit
models that are not linear models. Thus, although the terms "least squares" and "linear
model" are closely linked, they are not synonymous.
Describe the relationship between the five major crimes and the CCTV
installations and describe the relevance of data to the estimated regression model.
2.3 Artificial Neural Network
Artificial neural network is a statistical study algorithm modeled at the physical unit
of the brain, the physical unit of the brain. With the values of the neurons that have a
threshold and a function of each neuron, the value of each neuron is transmitted to the
following neurons to repeat the final output value to the next neuron. In other words,
the artificial neural network model has three levels of structure, which is output,
hidden, and input layer. It is utilized in research of artificial intelligence, such as
prediction and pattern recognition.
The goal of the neural network is to solve problems in the same way that the
human brain would, although several neural networks are much more abstract.
Modern neural network projects typically work with a few thousand to a few million
neural units and millions of connections, which is still several orders of magnitude
less complex than the human brain and closer to the computing power of a worm.
Fig. 1. Structure of Artificial Neural Network
Advanced Science and Technology Letters Vol.139 (FGCN 2016)
152 Copyright © 2016 SERSC
Page 4
3 Mail Title
3.1 Design of research
Based on the crime of murder, robbery, rape, theft and assault of the five major
crimes committed in 2011-2014, it is analyze the number of CCTV cameras for local
crime prevention.
As of [Figure 2], the number of crimes in 2011 was set as an independent variable.
The status of CCTV installation is set as a dependent variable. Analyze the correlation
between variables and derive linear relationship analysis to analyze the conformity of
the data. We propose a model for installing a CCTV camera to prevent violent crime
using artificial network.
Fig. 2. Crime status and CCTV installation status
Advanced Science and Technology Letters Vol.139 (FGCN 2016)
Copyright © 2016 SERSC 153
Page 5
Fig. 3. A model of CCTV Installation Status Using Artificial Neural Network
3.2. A Proposal of a CCTV Installation Model Using Artificial Neural Networks
When using Artificial Neural Network, the nine of hidden layer node models is
proposed for the most effective among the five cases. The number of hidden layer
nodes was divided by 1, 3, 5, 7, 9. The performance of the model assessed SSE, Steps
of training, and correctness. The correlation analysis enhanced the reliability of the
model. The figure below of [Figure 3] is a model with nine artificial neural network
nodes.
4 Experimental and verification
The correlation was analyzed to determine how much the number of crime related
crimes and the number of CCTV installations were related in 2011-2014. [Figure 4]
shows the visualization of correlation.
Fig. 4. The correlation between crime status and CCTV installation
Advanced Science and Technology Letters Vol.139 (FGCN 2016)
154 Copyright © 2016 SERSC
Page 6
The correlation between the five major crimes and the number of CCTV
installations has a strong linear correlation. Perform multiple linear regression
analysis to assess the conformity of the model. The following [Figure 5] shows the
result of multiple linear regression analysis.
Fig. 5. Multiple Linear Regression
The coefficient of determination describes the complete data as the estimated
regression model is closer to 1. The coefficient of determination is assumed to be
representative of the regression model estimated at 74.3 %.
Based on a multiple linear analysis of crime and CCTV, the use of Artificial
Neural Network is used to propose effective CCTV installation models for crime
crimes. The Artificial Neural Network has five properties and an output node that
predicts the installation of the CCTV and hidden node. The Artificial Neural Network
shows the weight for each of connection, the number of repetitions, the measurement
of the error level. The number of hidden layer nodes has been increased as a way to
improve the performance of the Artificial Neural Network. When the performance of
the hidden layer node is 9, the sum of the error is the smallest of the 0.0237. The
number of training steps has become a complex model with 413.
Hidden layer node = 1
Hidden layer node = 3
Advanced Science and Technology Letters Vol.139 (FGCN 2016)
Copyright © 2016 SERSC 155
Page 7
Hidden layer node = 5
Hidden layer node = 7
Hidden layer node = 9
Fig. 6. Artificial Neural Network model
Fig. 7. Accuracy of Artificial Neural Network
Performance evaluation of a model measures the correlation between forecasted
and actual values. The following [Figure 7] illustrates the correctness of each model.
The accuracy of the model indicates that the correlation between the actual values
and the predicted values is closer to 1, and that it is well predicted. Considering the
sum of the error and the number of precision of the discipline and the precision of the
Advanced Science and Technology Letters Vol.139 (FGCN 2016)
156 Copyright © 2016 SERSC
Page 8
model, the status of the CCTV installation for crime is proposed with 9 model
numbers per node.
5 Conclusion
In this thesis, we proposed a model for the installation of CCTV in accordance with
violent crime. To ease the public's anxiety and prevent crimes from spreading, the
public is installing CCTVs in the nation and the local governments as well as citizens
to prevent crime. CCTV plays a crucial role in preventing crime and securing
evidence at the same time.
In this thesis, it analyzes the relationship between crime and the number of CCTV
cameras, and uses the Artificial Neural Network to provide a model for crime
prevention for crime prevention.
By analyzing the spatial characteristics of crime zones, it will be possible to
propose effective CCTV installation models for crime prevention by proposing a
crime prevention model for crime zones.
References
1. Kim, S.: Crime and criminal policies in South Korea. Research Institute of Criminal
Policy Studies, South Korea, (2009)
2. Hyung Jo, i., Ku Lee, K.: Journal of Governmental Studies. vol. 18. 2012, (2012), pp. 187-
227
3. National Police Agency: Institute of Public Security Policy. Security Outlook, South
Korea, (2016)
4. Korea Research Institute for Human Settlements: A Study on the Urban Safety Methods
Considering the Spatial Characteristics of the Crime Occurrence. South Korea, (2014)
5. Korea Information Technology Promotion Agency: A Case Study on Four Social Safety
Net Sites Using Spatial Information. South Korea, (2012)
6. Jahan Hossain, Md. S. and Ahmad, Dr. N.: Artificial Intelligence Based Surface
Roughness Prediction Modeling for Three Dimensional End Milling. IJAST, vol. 45,
August, (2012), pp. 1-18
7. Ramchandra Baviskar, P. and Tungikar, V. B.: Multiple Cracks Assessment using Natural
Frequency Measurement and Prediction of Crack Properties by Artificial Neural Network.
IJAST, vol. 54, May, (2013), pp. 23-38.
8. Benacer I. and Dibi, Z.: Modeling and Simulation of Organic Field Effect Transistor
(OFET) Using Artificial Neural Networks. IJAST, vol. 66, May, (2014), pp. 79-88
Advanced Science and Technology Letters Vol.139 (FGCN 2016)
Copyright © 2016 SERSC 157