Data Science – deep learning of neural networks and their application in healthcare

Introduction: Artificial intelligence, which is a set of algorithms, currently does an impressive amount of work related to its analysis and processing. The use of the computing power of a large number of simple processors, as well as the compilation of a mathematical model for their joint operation based on the principle of organizing neural networks of cells of living organisms, constitutes an artificial neural network. Such a system is not programmed at the development stage into a final consumer product (as is usually the case, for example, with the software of a device), but «teaches» throughout its entire operation. «Teaching» is about finding the percentage relationship between neurons and input data, which ultimately leads to the identification of complex relationships between the provided data. These properties of training neural networks are already helping doctors in their work, making it easier and providing more readable data. Purpose of the study: to update information about the use of modern technologies for teaching neural networks in the healthcare sector. Tasks: to consider the terminology and designate technologies in Data Science used in healthcare; to find on peer-reviewed resources information about modern approaches to the analysis of accumulated information and present it in a public language; to demonstrate the advantages and disadvantages of using deep teaching of neural networks; detail the «future» of deep teaching of neural networks in healthcare. Results: a complex system of interconnection between neurons of a neural network with a correctly written program code, together with relevant and verified information, makes it possible to accurately find correlations of many statistical indicators in the field of healthcare. This fact will ultimately lead to improved medical care. A neural network can handle large amounts of information much faster and more accurately, which is a huge step towards personalized medicine. This became possible due to the accumulation of a sufficient amount of data in digital form, as well as the achievement of sufficient technical progress in the field of deep teaching of neural networks.