Sučasnì Informacìjnì Tehnologìï u Sferì Bezpeki ta Oboroni (Jan 2015)
THE MODEL OF DETECTING LEGISLATION VIOLATORS AT THE STATE BORDER USING FUZZY LOGIC INFERENCE
Abstract
The analysis showed the existing information and telecommunications systems are not able to make enough timely and quality support decision-making on criminal analysis and require further improvement. Application of fuzzy inference model to identify offenders legislation to control, according to the formulation of the mathematical problem of identifying offenders in a non-linear estimation of the object with the set of input variables and one output variable. For example, if the living condition is poor and the availability of appropriate vehicles is high, the degree of possibility of offense is high. Each sign is formalized as the level of compliance with the implementation of crime. Obviously, the higher the score, the more can be carried out illegal activities. Output variable is the level of the degree of possibility of the offense: transportation of contraband, weapons, drugs, etc. in a particular area of the border. In the fuzzy model of risk assessment of unlawful activity, all variables are presented as linguistic, universal set is measured on a scale of personnel based on their knowledge, experience and intelligence. Construction of membership functions of terms "low", "medium", "high", used to assess linguistic variables, by using the method of statistical analysis expert information. The contradiction between a large number of input variables and the number of rules is solved by constructing a hierarchical knowledge base. The method of activation used min. Since all the rules as a logical connection to of the conditions only applies fuzzy conjunction (operation "AND"). As a method of aggregation used operations min-conjunction. For accumulation endings rules applied max-disjunction. As a method defazzyfikation the method of center of gravity. Implementation of fuzzy inference model by using the package fuzzyTech. The practical implementation of this model was carried out in a software module "Identifying offenders" which proposed to include subsystems "risk" information and telecommunication systems "Hart" of the State Border Guard Service of Ukraine. The application of the model in contrast to the existing allows: use of quality indicators, taking into account inaccurate, approximate information about the meaning of the signs, the use of knowledge professionals from operational activities - experts who served in the form of fuzzy inference rules, more qualitative assessment of object investigated in criminal analysis.
