GRNN Approach Towards Missing Data Recovery Between IoT Systems

Author(s)
Ivan Izonin, Natalia Kryvinska, Pavlo Vitynskyi, Roman Tkachenko, Khrystyna Zub
Abstract

This paper describes the main reasons for the problem of filling missed values in IoT devices. The method of solving the missing data recovery task using the regression approach is proposed. It is based on the use of artificial intelligence tool. Authors describes the main procedures for applying GRNN to solving this task. A number of practical investigations were carried out on the restoration of missed data in the real environment monitoring dataset. For this purpose, the dataset for assessing the air quality in the city, which contains a lot of passes, was selected. The high accuracy of the proposed method was experimentally investigated. A comparison of the work of the neural network with a number of existing machine learning algorithms is carried out. It has been found that GRNN provides at least 5% higher accuracy compared to existing methods. Prospects for further research on GRNN modifications for improving the accuracy of its work are outlined.

Organisation(s)
Department of Marketing and International Business
External organisation(s)
Lviv Polytechnic National University
Pages
445-453
Publication date
2019
Peer reviewed
Yes
Austrian Fields of Science 2012
502050 Business informatics, 502052 Business administration
Keywords
Sustainable Development Goals
SDG 11 - Sustainable Cities and Communities
Portal url
https://ucrisportal.univie.ac.at/en/publications/e9c04d9f-05e9-41ce-bf43-280785fba8e6