Modeling and analysis of illegal immigrants in Malaysia using two-mode and scale-free networks

Abstract

Many complex systems in the form of evolving networked structures are omnipresent in various domain of life such as social, biological, ecological, technological, communications, transportations and in many others. These networks are very large and complex due to huge number of nodes and links. The advancements in technology and high computational power lead to profound impact in evaluating these complex networks empirically. Network analysis and modeling can be used to better understand the working mechanism of these highly heterogeneous and dynamic networks. The analysis and modeling of complex networks is a daunting task due to their structural complexity, structural dynamics, topological diversity, nodal diversity and dynamical complexity. For decades the ”Erdos-Renyi” random model has been a guide for research until the recently proposed scale-free network model. Recent researches in the field of complex networks is focused on the small-world effect and scale-free phenomenon which proves that the degrees are size independent and it follows power-law behavior. There are a large number of illegal immigrants in Malaysia. The number keeps rising and according to estimation; there are approximately 2 million illegal immigrants from various countries of the world. The distribution of illegal immigrants found throughout Malaysia, is inhomogeneous. In some states number of illegal immigrants is significantly higher than others. Having great number of illegal immigrants creates law and order challenges for law enforcing agencies, such as Royal Police of Malaysia. In this research we have formalized the distribution of illegal immigrants convicted in different crimes (2007 to 2012) into weighted two-mode complex network. With the help of graph theory, we chose several important network analysis metrics such as clustering coefficient, nodes strength distribution, betweenness, closeness, density, path-length etc. to analyze the behavior of this socio-geographic network in terms of important nodes, links and overall topology. This research has found that network of illegal immigrants has self-organized itself into scale-free network and therefore behaves robustly under random attack of links removal. Due to scale-free topology this network has shown highly clustered behavior between nodes. Further, majority of illegal immigrants are convicted from few countries in few specific states of Malaysia and mostly from neighboring countries. Moreover, types of crime committed by immigrants have shown the power-law distribution, which clearly indicates that few specific crime types have been committed many times by immigrants. Finally, we found that the random searching of illegal immigrants by police cannot minimize this problem and targeted approach may bring better results as the network has self-organized into scale-free topology. The results of network analysis satisfactorily support our hypothesis that the illegal immigrants distribution phenomena is very far from random network.