Open AccessA System for Detecting Anomalous Observations in Data on Healthcare Services
Author(s) -
Semen Blyumin,
R. V. Scheglevatykh,
Aleksandar Naydenov,
Anton Sysoev
Publication year - 2021
Publication title -
vestnik tambovskogo gosudarstvennogo tehničeskogo universiteta
Language(s) - English
Resource type - Journals
eISSN - 2542-1409
pISSN - 0136-5835
DOI - 10.17277/vestnik.2021.03.pp.356-367
Subject(s) - artificial neural network , computer science , data mining , classifier (uml) , anomaly detection , artificial intelligence , machine learning , software , metric (unit) , process (computing) , engineering , operations management , programming language , operating system
A description of the mathematical model of a neural network classifier of data on healthcare in the institutions of the Lipetsk region is given in order to identify atypical (abnormal) records. Anomaly detection refers to the problem of finding data that is inconsistent with some expected process behavior or metric occurring in the system. Due to the large number of inputs to the neural network model, the time it takes to process the incoming information also increases. To assess what factors should be transmitted to the input of the neural network classifier, an approach to the reduction of the neural network model based on sensitivity analysis is proposed. The description of a set of software tools for solving the problem is presented.