The aim of the study was to assess biomedical risk factors for preeclampsia in pregnant women with chronic arterial hypertension (CAH) and on this basis to create the neural network system for calculating the probability of developing preeclampsia in these women. Materials and Methods. Pregnancy and delivery outcomes were analyzed in 548 patients with pre-existing arterial hypertension (AH): 318 with CAH and 230 with preeclampsia secondary to CAH. Risk factors were calculated using the OpenEpi program (UK). A combined method of global optimization and neural network method of information compression were used when training the developed neural network system. Results. There were identified the main risk factors for developing preeclampsia in pregnant women with CAH: hereditary burden of hypertension; hypertensive disorders in previous pregnancies; hypertension during more than five years; the initial diastolic blood pressure being more than 80 mm Hg; body mass index more than 30; tobacco smoking; nulliparity; chronic pyelonephritis and gastritis; hypertensive disease stage II; degree II and III AH; hypertensive retinal angiopathy; left ventricular hypertrophy; lack of regular antihypertensive therapy before and during pregnancy; late treatment initiation. The data obtained were used to train and test the neural network software and to develop the "Neuro_Chronic-neural network system for predicting secondary preeclampsia in pregnant women with chronic arterial hypertension". The system includes two modules. The first module is designed to train the neural network software model using a given set of images, the second module provides evaluation of preeclampsia developing during pregnancy in a particular patient in the form of five probability options-from very low to very high-after entering the parameters obtained during the anamnestic and clinical examination into the corresponding fields. Conclusion. Revealing the proposed predictors of preeclampsia in pregnant women with CAH and entering these data into the developed computer program will enable physicians to determine the probability of preeclampsia developing during gestation at the outpatient stage and to take timely preventive measures in pregnant women at high-risk.
The evaluation of forest fire danger is the actual issue. We see the improving possibility of forest fire danger evaluation. It means the correct definition of places with high probability of ignition occurrence. To solve this problem we developed the new computational scheme, which considers the variety of different factors. The offered method is based on the neural network technology and consists of three main blocks. 1. Block of forest combustible material humidity and soil humidity. It calculates the danger on the base of soil moisture content, averaged humidity of dead wood and forest inflammability according to Nesterov's criterion. 2. Block of human factor and thunderstorm activity. It considers the closeness of localities, highways, the presence of major hazard and thunderstorm activity. 3. Data base of wildfire hazard categories on the defined area. Forest sites are divided into five categories of wildfire hazard according to the composition of vegetation, topography of the area and its condition.Our research group also established the following developments in the field of forest fires: Mathematical modeling of forest fire distribution (Original equations, the splitting into physical processes, regularization according to the method of Buleev, Scalar sweep, the combination of the computational grid with a geographical map); Electronic forester (device for gathering information about the state of forests, needed to assess the likelihood of fire and fire modeling);
This article considers the questions connected with creation of optimum algorithms using the laws of thermodynamics as applied to a computing process. Ideas and methods of phenomenological and statistic thermodynamics are used to estimate the amount of calculations or volume of the neural network. Introduction of the other thermodynamic functions, besides entropy, and also defi nition of the three thermodynamic origins in the context of calculations allow to study stability, organize the parameters according their information weights, carry out the decomposition of the complex systems, construct the rapid algorithms. The way of creation of the neural network structure is offered, consisting in use of the pre-trained fragments.
Состояние вопроса. Одной из главных проблем при построении математических моделей турбулентных гидродинамических сред остается высокая вычислительная сложность компьютерных расчетов, поскольку требуется получить решение нестационарной задачи с сеткой такого пространственного шага, чтобы он соответствовал размерам самых малых вихревых структур. Существуют рекомендации по использованию разностных вычислительных схем для турбулентной динамики жидкости, полученные при моделировании розлива нефтепродуктов по водной поверхности. В связи с этим актуальным является получение устойчивого расчета математической модели динамики вязкой несжимаемой жидкости согласно описанию Эйлера с учетом влияния турбулизации и ускорение расчетов в параллельном интерфейсе CUDA. Материалы и методы. При проведении вычислительных экспериментов применяются методы математического моделирования физических объектов, метод численного интегрирования дифференциальных уравнений, метод противоточных производных для повышения устойчивости разностных схем. Для оценки турбулентной вязкости сплошной среды используется метод Секундова. Результаты. Создана устойчивая параллельная реализация математической модели, описывающей процессы в сплошной среде с учетом влияния турбулизирующих гидродинамических структур, устойчивость которой достигается за счет применения метода противоточных производных и замены традиционных разностных схем при расчете полей скорости и давления на четырехточечные аналоги. Предложен метод равномерного распределения вычислительных ресурсов графического ускорителя при больших размерах сетки. Получен устойчивый расчет модели динамики сплошной среды с вычислением турбулентной вязкости по модели Секундова на значительном временном отрезке. Выводы. Построенное решение на основе интерфейса CUDA позволяет достичь ускорения вычислений от 2 до 8 раз в зависимости от возможностей аппаратной части. Разработанная система может являться инструментом исследования объектов энергетической отрасли, для которых требуется ускорение принятия управляющих решений по сравнению со стандартным программным обеспечением, не предусматривающим распараллеливания расчетов. Достоверность результатов вызвана соответствием системных условий конфигурациям двухфазных систем питания энергетических установок.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.