19 research outputs found
Development and experimental study of an intelligent water quality monitoring system based on the internet of things
No full textThe goal of this work is to create an intelligent internet of things (IoT)-based water quality monitoring system that will effectively monitor and analyze water parameters, collect real-time data, and provide critical information for decision-making in water management and environmental issues. Provide data transfer over wireless networks such as Wi-Fi or Bluetooth. The scientific novelty of the project lies in the development of an innovative system that combines modern IoT technologies and machine learning methods to provide comprehensive and accurate water quality monitoring, which is a significant contribution to water management and environmental safety. Five sensors are connected to Arduino-Mega 2560, ESP-32-E in a discrete manner to determine water parameters. The extracted sensor data is transferred to a desktop application developed on the Blynk App platform and compared with World Health Organization (WHO) standard values. Based on the measurement results, the proposed system can successfully analyze water parameters using the fast forest binary classifier to determine whether the tested water sample is potable or not. An intuitive user interface has been created that will allow users to monitor and analyze water quality data in real time. Provide the ability to create graphs, charts, and reports for visual presentation of data
Using Machine Learning Algorithm for Diagnosis of Stomach Disorders
No full textMedicine is one of the rich sources of data, generating and storing massive data, begin from description of clinical symptoms and end by different types of biochemical data and images from devices. Manual search and detecting biomedical patterns is complicated task from massive data. Data mining can improve the process of detecting patterns. Stomach disorders are the most common disorders that affect over 60% of the human population. In this work, the classification performance of four non-linear supervised learning algorithms i.e. Logit, K-Nearest Neighbour, XGBoost and LightGBM for five types of stomach disorders are compared and discussed. The objectives of this research are to find trends of using or improvements of machine learning algorithms for detecting symptoms of stomach disorders, to research problems of using machine learning algorithms for detecting stomach disorders. Bayesian optimization is considered to find optimal hyperparameters in the algorithms, which is faster than the grid search method. Results of the research show algorithms that base on gradient boosting technique (XGBoost and LightGBM) gets better accuracy more 95% on the test dataset. For diagnostic and confirmation of diseases need to improve accuracy, in the article, we propose to use optimization methods for accuracy improvement with using machine learning algorithms
A Comparison of Machine Learning Algorithms in Predicting Lithofacies: Case Studies from Norway and Kazakhstan
No full textDefining distinctive areas of the physical properties of rocks plays an important role in reservoir evaluation and hydrocarbon production as core data are challenging to obtain from all wells. In this work, we study the evaluation of lithofacies values using the machine learning algorithms in the determination of classification from various well log data of Kazakhstan and Norway. We also use the wavelet-transformed data in machine learning algorithms to identify geological properties from the well log data. Numerical results are presented for the multiple oil and gas reservoir data which contain more than 90 released wells from Norway and 10 wells from the Kazakhstan field. We have compared the the machine learning algorithms including KNN, Decision Tree, Random Forest, XGBoost, and LightGBM. The evaluation of the model score is conducted by using metrics such as accuracy, Hamming loss, and penalty matrix. In addition, the influence of the dataset features on the prediction is investigated using the machine learning algorithms. The result of research shows that the Random Forest model has the best score among considered algorithms. In addition, the results are consistent with outcome of the SHapley Additive exPlanations (SHAP) framework
Dynamic Simulation of a Solar Hot Water Heating System for Kazakhstan Climate Conditions
No full textMATHEMATICAL MODELING OF WATER MOVEMENT DURING A DAM BREAK USING THE VOF METHOD
No full textRiver valleys in mountainous areas are often subject to heavy rains and melting glaciers, resulting in the risk of mudflows and the destruction of hydraulic protective structures. In order to minimize the potential risk and negative outcomes of a disaster, both on an individual and environmental scale, it is crucial to possess essential information. This includes understanding the timing, location, and extent of flooding, as well as comprehending the force of water flow impact on protective structures. In the research, the numerical process of the movement of the water flow caused by the breakthrough of the dam is investigated. A two-dimensional numerical model of water flow during a dam break was constructed using the VOF method to describe the described process. With the help of the VOF method, the movement of the water surface is captured, while maintaining the law of conservation of mass. The mathematical model consists of Reynolds-averaged incompressible Navier-Stokes equations and includes the interphase equation. The turbulent k-e model was used to close the system of equations. The numerical algorithm used is PISO (Pressure-Implicit with Splitting of Operators). The obtained numerical results agree with the experimental data, indicating the developed algorithm’s reliability and accuracy. The results are presented as comparative graphs and images showing the contour of the free surface movement along the experimental reservoir. A numerical model that has been tested in this way can provide significant support in preventing the devastating consequences of a dam break and providing timely assistance during the evacuation of the population
Machine learning algorithms for stratigraphy classification on uranium deposits
No full textMachine learning today becomes more and more effective instrument to solve many particular problems, where there are difficulties to apply well known and described math model. In other words - it is a great tool to describe non-linear phenomena. We tried to use this technique to improve existing process of stratigraphy, and reduce costs on site by applying computer leaded predictions on the basis of existing on-field collected data. Article describes usage of machine learning algorithms for stratigraphy boundaries classification based on geophysics logging data for uranium deposit in Kazakhstan. Correct marking of stratigraphy from geophysics logging data is complex non-linear task. To solve this task we applied several algorithms of machine learning: random forest, logistic regression, gradient boosting, k nearest neighbour and XGBoost
Порівняння згорткових нейронних мереж для розпізнавання казахської жестової мови
No full textFor people with disabilities, sign language is the most important means of communication. Therefore, more and more authors of various papers and scientists around the world are proposing solutions to use intelligent hand gesture recognition systems. Such a system is aimed not only for those who wish to understand a sign language, but also speak using gesture recognition software. In this paper, a new benchmark dataset for Kazakh fingerspelling, able to train deep neural networks, is introduced. The dataset contains more than 10122 gesture samples for 42 alphabets. The alphabet has its own peculiarities as some characters are shown in motion, which may influence sign recognition.
Research and analysis of convolutional neural networks, comparison, testing, results and analysis of LeNet, AlexNet, ResNet and EffectiveNet – EfficientNetB7 methods are described in the paper. EffectiveNet architecture is state-of-the-art (SOTA) and is supposed to be a new one compared to other architectures under consideration. On this dataset, we showed that the LeNet and EffectiveNet networks outperform other competing algorithms. Moreover, EffectiveNet can achieve state-of-the-art performance on nother hand gesture datasets.
The architecture and operation principle of these algorithms reflect the effectiveness of their application in sign language recognition. The evaluation of the CNN model score is conducted by using the accuracy and penalty matrix. During training epochs, LeNet and EffectiveNet showed better results: accuracy and loss function had similar and close trends. The results of EffectiveNet were explained by the tools of the SHapley Additive exPlanations (SHAP) framework. SHAP explored the model to detect complex relationships between features in the images. Focusing on the SHAP tool may help to further improve the accuracy of the modelДля людей с ограниченными возможностями жестовый язык является важнейшим средством общения. Поэтому все больше авторов различных работ и ученых по всему миру предлагают решения для использования интеллектуальных систем распознавания жестов рук. Такая система предназначена не только для тех, кто хочет понимать жестовый язык, но и говорить с помощью программного обеспечения для распознавания жестов. В данной работе представлен новый эталонный набор данных для дактильного алфавита казахского языка, способный обучать глубокие нейронные сети. Набор данных содержит более 10122 образцов жестов для 42 алфавитов. Алфавит имеет свои особенности, так как некоторые символы показаны в движении, что может влиять на распознавание жестов.
В статье описаны исследования и анализ сверточных нейронных сетей, сравнение, тестирование, результаты и анализ методов LeNet, AlexNet, ResNet и Effectivenet – EfficientNetB7. Архитектура EffectiveNet является самой современной и новой по сравнению с другими рассматриваемыми архитектурами. На этом наборе данных мы показали, что сети LeNet и EffectiveNet превосходят другие конкурирующие алгоритмы. Кроме того, EffectiveNet обеспечивает высочайшую производительность на других наборах данных жестов.
Архитектура и принцип работы этих алгоритмов отражают эффективность их применения при распознавании жестового языка. Оценка модели СНС проводится с использованием матрицы точности и штрафов. В периоды обучения LeNet и EffectiveNet показали лучшие результаты: функции точности и потерь имели схожие и близкие тенденции. Результаты EffectiveNet были объяснены с помощью инструментов структуры Аддитивных объяснений Шепли (SHAP). С использованием SHAP исследовалась модель для обнаружения сложных взаимосвязей между элементами изображений. Сосредоточение внимания на инструменте SHAP может помочь еще больше повысить точность моделиДля людей з обмеженими можливостями жестова мова є найважливішим засобом спілкування. Тому все більше авторів різних робіт і вчених по всьому світу пропонують рішення для використання інтелектуальних систем розпізнавання жестів рук. Така система призначена не тільки для тих, хто хоче розуміти жестову мову, а й говорити за допомогою програмного забезпечення для розпізнавання жестів. У даній роботі представлений новий еталонний набір даних для дактильного алфавіту казахської мови, здатний навчати глибокі нейронні мережі. Набір даних містить більше 10122 зразків жестів для 42 алфавітів. Алфавіт має свої особливості, так як деякі символи показані в русі, що може впливати на розпізнавання жестів.
У статті описано дослідження та аналіз згорткових нейронних мереж, порівняння, тестування, результати та аналіз методів LeNet, AlexNet, ResNet та Effectivenet – EfficientNetB7. Архітектура EffectiveNet є найсучаснішою і новою в порівнянні з іншими розглянутими архітектурами. На цьому наборі даних ми показали, що мережі LeNet та EffectiveNet перевершують інші конкуруючі алгоритми. Крім того, EffectiveNet забезпечує найвищу продуктивність на інших наборах даних жестів.
Архітектура і принцип роботи цих алгоритмів відображають ефективність їх застосування при розпізнаванні жестової мови. Оцінка моделі ЗНМ проводиться з використанням матриці точності і штрафів. У періоди навчання LeNet і EffectiveNet показали кращі результати: функції точності і втрат мали схожі і близькі тенденції. Результати EffectiveNet були пояснені за допомогою інструментів структури адитивних пояснень Шеплі (SHAP). З використанням SHAP досліджувалася модель для виявлення складних взаємозв'язків між елементами зображень. Зосередження уваги на інструменті SHAP може допомогти ще більше підвищити точність модел
Soil Salinity Classification Using Machine Learning Algorithms and Radar Data in the Case from the South of Kazakhstan
No full textSoil salinity is one of the major impact factors on agriculture in the South of Kazakhstan. Prediction and estimation of soil salinity before planting a season usually helps to plan for the leaching of the salt. In the paper, satellite data such as radar data and machine learning algorithms, were used to classify soil salinity. Numerical results were presented for the Turkestan region, which contains more than 102 points. The machine learning algorithms, including Gaussian Process, Decision Tree, and Random Forest, were compared. The evaluation of the model score was realized by using metrics, such as accuracy, Recall, and f1. In addition, the influence of the dataset features on the classification was investigated using machine learning algorithms. The research results showed that the Gaussian Process model has the best score among considered algorithms. In addition, the results are consistent with the outcome of the Shapley Additive exPlanations (SHAP) framework
Simulating Fracture Networks Using Machine Learning Approaches for Geological Data from Kazakhstan
No full textPoster shows the fracture network model in the subsurface using mathematical analysis including geostatistical analysis and/or deep learning algorithms. Thus, the efficient fracture model verified with the results of the Kazakhstan data was investigated to enrich the subsurface characterization.</p
