Methods and algorithms for storage, systematization and processing of medical electroencephalographic data

Alekseev Dmitry Mikhailovich, Minyuk Anatoly Nikolaevich, Shumilin Aleksandr Sergeyevich, Ponimash Zachary Alekseevich

the purpose of the research is to develop and evaluate the effectiveness of the overall scheme of the cloud platform, which provides a certain range of tasks, as well as the selection and analysis of the classifier to solve the problem of the allocation of areas of epileptic activity as part of the cloud platform. As a result of the research, a scheme of cloud platform for storage, systematization and processing of medical data has been developed, which includes four main levels. Separation of data flows into levels, standardization of information transfer protocols and formats of its storage allow to create a flexible, universal and highly reliable medical information system. The architecture allows you to quickly integrate into existing medical systems and implement used in medical practice hardware, as well as individual mobile devices. A single space for storing impersonal data makes it possible to carry out research of a significant array of classified medical information by means of machine learning. Within the framework of the developed cloud platform for storage, systematization and processing of medical data, the effective work of the classifier based on the identification of potentially dangerous amplitudes has been achieved. Accuracy (93%) and high speed of this classifier are optimal for the developed cloud platform.

Keywords:cloud computing, classifier, medical research, machine learning, data storage, data processing, epileptic activity, data systematization