Real Statistics Help To Reach Criminals Via Fuzzy MCDM Approach

Detection of criminals is quite a particular challenge. Concerned departments have limited resources that can be assigned to search for criminals, which can be an issue when time is of the essence. There are two tracks that can give answers to this problem: behavioural and situational similarities. The first aspect relies on the behaviour of the offender, described by particular methods of crime. Understanding the workings of the human psyche, and treating and modifying aberrant behaviour is the nucleus of the dynamic field of Behaviour Analysis. The second is based on spatio-temporal similarities: analyses are conducted in accordance with crime mapping theories. Law enforcement agencies monitor criminal incidents. With additional geographic and demographic data, law enforcement analysts look for spatio-temporal patterns in these incidents in order to either predict future criminal activity or to find current criminal activity in light of past incidents. However, a third aspect can be considered: forensic cased data (such as DNA, shoe marks or earmarks) which is also considered very vital in this process.

CCTV footages help a lot these days. Few local and conventional methods are also utilized. This study will attempt to combine behavioural, situational, and forensic information for real statistical data based on different techniques used and employed in different movies. Covering these three aspects under the light of modern tools used in movies is a way to retrieve data and then combine techniques for crimes analysis. To do so, several mathematical methods that effectively compute these similarities can be put altogether. Our approach is based on a fuzzy multi-criteria decision making (MCDM) method in order to deal with that issue. Indeed, in a crime analysis context, fuzziness permits to model experts’ experience and handle vagueness whereas MCDM is useful to evaluate and combine similarities stemming from multiple criteria. Zainab murder case is taken as preliminary data. Hard procedure adopted by police led the case with very slow speed. Even CCTV footages were not enough to identify the suspect. By using behavioural and situational similarities locals had suspected a man long before police had any clue of it. Police gathered DNA information and through rigorous process arrested the criminal.

The process took more than 20 days. If police had went through combined behavioural, situational and forensic details at once through proper planning or by using any mathematical tools, suspect would have arrested within a week. Limitations to this research is capturing true statistics from concerned departments about different incidents actually happened in vague and fuzzy environment. With the potential to benefit to all police organizations, this research will lay the ground for conducting new experiments and sharing results in quick criminal detection.