APPROACH TO MODELLING AND ACCOUNTING FOR LATENT VALUES OF LEGAL STATISTICS INDICATORS

The aim of the work is to describe, based on using fuzzy set theory methods, a new approach to modelling and accounting for latent values of legal statistics indicators used in analytical work by law enforcement officers. The reason for the emergence of latent values is that due to the peculiarities of the subject area, there is partial inconsistency and incompleteness of official statistical data. The choice of mathematical method is determined by the subjective nature of information about latent values, which can be obtained in the form of expert assessments as a result of criminological studies. Using two types of LR membership functions for fuzzy estimates depending on the types of indicators is justified. The work develops a comparison method of the two types of estimates obtained with target estimates of legal statistics indicators, for which the use of fuzzy estimates with LPR membership functions was previously justified; establishes correspondence as the intersection of these fuzzy estimates. As a scalar characteristic of the correspondence estimate, the paper proposes either the normalized value of the membership function area or the median estimate of this membership function, gives a numerical example. In addition, a method for obtaining estimates of correspondence of a set of legal statistics indicators to their target values is described. Law enforcement officers, can use the results obtained, for example, in analytical work to assess the operational situation in a particular region, and they allow for increased validity of managerial decisions.