Neural network model of human intoxication functional state determining in some problems of transport safety solution

Abstract

Ключевые слова: нейросетевая модель, пупиллометрия, зрачковая реакция, предрейсовый медицинский осмотр, функциональное состояние опьянения водителя, принятие решений Работа выполнена при финансовой поддержке Министерства образования и науки РФ в рамках государственного задания по проекту № 8.6141.2017/8.9.

The Technology of Collecting Initial Data for Constructing Models for Assessing the Functional State of a Person by Pupillary Response to Changes In Illumination

“…The effectiveness of the developed technology and HSC is shown on the initial data of the human intoxication FS. To compare the results obtained, we used the previously constructed models (model of a single-layer perceptron, model of a collective of neural networks) [4,15,18], where the initial data for training and testing were obtained by computer vision methods. Similarly, according to the topology and construction principles, we formed the models, the initial data for which were obtained using the given technology.…”

“…With the development of technical means of computer vision, the use of the pupillometry method in assessing functional states seems to be highly promising not only in medical diagnostics systems, but also in public and transport security systems [4], as well as life and labor safety [5,6]. In particular, the improvement of computer vision methods makes it possible to move away from the use of video pupillograph devices, which are usual for medical institutions, as well as use the inexpensive high-resolution video cameras and the corresponding software to assess functional states [7,8].…”

Identification of Bots in Social Networks based on Data Mining Technologies

“…In this work, a neural network, a decision tree, and linear regression are used as tools for constructing bots identification models in social networks [17][18][19]. The choice of these methods is due to their high efficiency in solving problems of diagnostics, pattern recognition, objects classification, and their condition assessment [20][21][22][23].…”

Neuro-fuzzy model of fuzzy rules formation for objects state evaluation in conditions of uncertainty

“…Для учета лингвистической неопределенности и неполноты данных целесообразно использовать методы нечеткой логики и алгоритмы нечеткого логического вывода. Теория нечетких множеств [Zadeh, 1965], как одно из направлений искусственного интеллекта [Akhmetvaleev, Katasev, 2018;Ge et al, 2017;Guskov et al, 2018;Ismagilov et al, 2018], позволяет строить нечеткие модели объектов с использованием лингвистических переменных и механизмов логического вывода. При этом нечеткая модель представляет систему нечетко-продукционных правил и алгоритм вывода на правилах.…”