A Regularization-Based Method of Identification of Information Objects
Sergey I. Suyatinov, A. M. Khudyakov, M. S. Uvarova · Automatic Documentation and Mathematical Linguistics · 2022
This article considers the problem of identification of information objects whose attributes are represented by number and character sequences. It is shown that this problem is classified as ill-posed. A regularization method based on including a priori information about the probability of errors in descriptions of attributes of identification objects is proposed. Examples of the application of the proposed method to the problem of identification of individuals based on personal data are given. Tables of estimated error probabilities are compiled using statistical methods.