THE USE OF A DATA-EXCITATION ALGORITHM TO PROTECT CONFIDENTIALITY

Abstract

Combining a large number of different technologies, such as the Internet of Things, cloud computing, computational computing and machine learning, contributes to the rapid and active spread of technological development in various fields, such as health, energy, agriculture, transportation, etc. The increase in the number of publicly available gadgets has contributed to the rapid growth of the Internet of Things, becoming one of the main sources of large data flows.

Cyberspace covers not only the physical sphere, but also the human, a huge amount of information (data) becomes available for analysis. Analytical processing of large amounts of data, with the development of their generation speed, gives excellent results and provides extreme accuracy in creating important ideas. One of the approaches that attracts the most attention is in-depth training, which provides high performance with large amounts of data. Various areas, including the above, work closely with data privacy. They show a tendency to increase the consequences due to the disclosure of confidential data to third parties, attacks on databases and more. Corporations and other organizations are constantly striving to ensure maximum confidentiality and that all information is stored on the company's local servers. To prevent breaches of privacy, machine learning, together with data analytics, should implement all possible privacy protection scenarios to ensure that users' privacy is not compromised. There are many approaches to maintaining confidentiality, however, the sheer size of large amounts of data and data flows make maintaining a confidentiality a challenge. The main problem among the existing approaches is their inability to maintain the right balance between confidentiality, usefulness and efficiency when dealing with large amounts of data. Some effective approaches provide good privacy but do not provide sufficient performance during data operations, while others, on the contrary, provide good performance but do not provide a high level of privacy.

This paper investigates privacy protection algorithms for large-scale data based on machine learning. In today's world in everyday life and for researchers and practitioners, the problem of protecting the privacy of big data is the most basic problem and urgent challenge. Based on the results of this research, a privacy protection method has been improved for more efficient use of energy. The method has also been further developed to enable its implementation in the healthcare system.