The Development and Implementation of the Data-Binding Algorithm in the State Civil Information System

Abstract

This paper describes the development of a data-binding algorithm. The algorithm is based on a variant of the adaptive record similarity measure (ARSM). Optimization of this variant for solving the problem of data binding is carried out using a machine-learning algorithm. The approach to estimating the adequacy of ARSM for data binding consists in introduction and automated selection of the threshold values by the machine-learning algorithm, which is achieved by introducing the fitness function representing the modification F_measure.