Optimization of Pathology Diagnosis Business Process by Applying Chaos Theory and Fractal Analysis to EEG-signal Processing

By 2018, there are more than 70 million people suffering from various forms of one of the most common neurological diseases – epilepsy. Nowadays, most of the patients do not have the ability to predict the onset of an attack in advance, which is due to complex symptoms that are difficult to predict. This research paper considers the problem of epileptic seizures prediction from the point of view of the theory of nonlinear dynamical systems. The aim of the work is to test the hypothesis of the data compliance of neural networks of the cerebral cortex to multifractal nature. The main objectives of the study are to analyze and build a prognostic model for detecting the initial stage of an attack and to develop a set of recommendations for improving existing implantable devices that solve the described problem. The result of this work is the extension of the existing model based on the support vector machine through the analytical approach RQA.