Network Machine Research Articles

Bank Locker Security System Using Machine Learning with Face and Liveness Detection

To increase reliability of face recognition system, the system must be able to distinguish real face from a copy of face such as a photograph. This system introduces a distinguished approach towards Machine learning and Neural Network, in which we propose a fast and memory efficient method of live face detection for embedded face recognition system, based on the analysis of the movement of the eyes. We detect eyes in sequential input images and calculate variation of each eye region to determine whether the input face is a real face or not.

Predicting the impacts of land use/land cover changes on seasonal urban thermal characteristics using machine learning algorithms

Changes in land use/land cover (LULC) and land surface temperatures (LST) contribute significantly to the formation and intensity of urban heat islands (UHI) effects. The urban thermal field variance index (UTFVI) can effectively describe any city's UHI (thermal characteristics) effect. This study aims to assess and predict the seasonal (summer and winter) UTFVI scenario to evaluate the thermal characteristics of Sylhet city, Bangladesh. Landsat 4–5 TM and 8 OLI images from 1995 to 2020 were used to assess the previous status of LULC and UTFVI and predict the future changes for 2025 and 2030 using cellular automata and artificial neural network machine learning algorithms. Prediction results indicate a substantial increase in urban built-up areas by 42% and 44% in 2025 and 2035, followed by reductions in green cover (21% and 22%), bare land (20% and 21%) and water bodies (1%). The rapid expansion of built-up areas will lead to 13 km2 and 14 km2 stronger UTFVI zones in the predicted years. The study provides effective strategies for mitigating the UTFVI effects by avoiding dense infrastructural development, increasing plantation and water bodies, rooftop gardening and using white colour roofs in construction. The findings of this study will allow the urban planners, policymakers and local government to ensure an eco-friendly, inclusive and sustainable urban development through functional modification and replacement of the LULC distribution depending on the present and future circumstances.

Prediction of metabolic and pre-metabolic syndromes using machine learning models with anthropometric, lifestyle, and biochemical factors from a middle-aged population in Korea

BackgroundMetabolic syndrome (MetS) is a complex condition that appears as a cluster of metabolic abnormalities, and is closely associated with the prevalence of various diseases. Early prediction of the risk of MetS in the middle-aged population provides greater benefits for cardiovascular disease-related health outcomes. This study aimed to apply the latest machine learning techniques to find the optimal MetS prediction model for the middle-aged Korean population.MethodsWe retrieved 20 data types from the Korean Medicine Daejeon Citizen Cohort, a cohort study on a community-based population of adults aged 30–55 years. The data included sex, age, anthropometric data, lifestyle-related data, and blood indicators of 1991 individuals. Participants satisfying two (pre-MetS) or ≥ 3 (MetS) of the five NECP-ATP III criteria were included in the MetS group. MetS prediction used nine machine learning models based on the following algorithms: Decision tree, Gaussian Naïve Bayes, K-nearest neighbor, eXtreme gradient boosting (XGBoost), random forest, logistic regression, support vector machine, multi-layer perceptron, and 1D convolutional neural network. All analyses were performed by sequentially inputting the features in three steps according to their characteristics. The models’ performances were compared after applying the synthetic minority oversampling technique (SMOTE) to resolve data imbalance.ResultsMetS was detected in 33.85% of the subjects. Among the MetS prediction models, the tree-based random forest and XGBoost models showed the best performance, which improved with the number of features used. As a measure of the models’ performance, the area under the receiver operating characteristic curve (AUC) increased by up to 0.091 when the SMOTE was applied, with XGBoost showing the highest AUC of 0.851. Body mass index and waist-to-hip ratio were identified as the most important features in the MetS prediction models for this population.ConclusionsTree-based machine learning models were useful in identifying MetS with high accuracy in middle-aged Koreans. Early diagnosis of MetS is important and requires a multidimensional approach that includes self-administered questionnaire, anthropometric, and biochemical measurements.

Identification of hepatic steatosis in living liver donors by machine learning models.

Selecting an optimal donor for living donor liver transplantation is crucial for the safety of both the donor and recipient, and hepatic steatosis is an important consideration. We aimed to build a prediction model with noninvasive variables to evaluate macrovesicular steatosis in potential donors by using various prediction models. The study population comprised potential living donors who had undergone donation workup, including percutaneous liver biopsy, in the Republic of Korea between 2016 and 2019. Meaningful macrovesicular hepatic steatosis was defined as >5%. Whole data were divided into training (70.5%) and test (29.5%) data sets based on the date of liver biopsy. Random forest, support vector machine, regularized discriminant analysis, mixture discriminant analysis, flexible discriminant analysis, and deep neural network machine learning methods as well as traditional logistic regression were employed. The mean patient age was 31.4 years, and 66.3% of the patients were men. Of the 1652 patients, 518 (31.4%) had >5% macrovesicular steatosis on the liver biopsy specimen. The logistic model had the best prediction power and prediction performances with an accuracy of 80.0% and 80.9% in the training and test data sets, respectively. A cut‐off value of 31.1% for the predicted risk of hepatic steatosis was selected with a sensitivity of 77.7% and specificity of 81.0%. We have provided our model on the website (https://hanseungbong.shinyapps.io/shiny_app_up/) under the name DONATION Model. Our algorithm to predict macrovesicular steatosis using routine parameters is beneficial for identifying optimal potential living donors by avoiding superfluous liver biopsy results.

Multifunctional hydrogel as wound dressing for intelligent wound monitoring

• The hydrogel exhibits antibacterial, hemostatic and adhesive properties. • The hydrogel dressing monitors the wound pH in real-time through colorimetry. • The whole intelligent wound monitoring process mainly include three parts. • A combination of multifunctional hydrogel dressing and machine learning technology. Chronic wound remains a significant challenge in clinical care due to the long treatment process and the poor quality of tissue regeneration affected by bacterial infection, calling timely and effective wound management approach to achieve real-time intelligent wound monitoring and promote wound healing. In this work, a multifunctional hydrogel is employed as wound dressing for intelligent wound monitoring, which not only have the functions of antibacterial, hemostatic and adhesive properties for effectively promoting wound healing, but also can realize real-time wound status monitoring (e.g., pH). The whole intelligent wound monitoring process mainly includes three parts: wound recognition, real-time status monitoring and personalized wound management. Among them, the customized hydrogel wound dressing can accurately match the contour of the wound for precise treatment. Moreover, the personalized wound management model with high accuracy (94.47%) based on the convolutional neural network (CNN) machine learning algorithm can analyze and evaluate the wound healing and infection status through the colorimetric signal of the hydrogel dressing. This multifunctional hydrogel wound dressing, which integrates precise treatment, real-time monitoring and personalized management for intelligent wound monitoring, provides an advanced solution to accelerate wound healing and reduce bacterial infections and plays an inestimable step for future intelligent wound management.

Transformer Fault Diagnosis Method Based on Neural Network and D-S Evidence Theory

Abstract The BP neural network is used to calculate the basic fault probability distribution of the dissolved gas in the transformer oil and the core grounding online monitoring data, and the Dempster-Shafer evidence theory is used to fuse the multi-source information of the basic probability of various types of faults to obtain a transformer fault diagnosis model. Transformer samples are used to verify the model, and the support vector machine and convolutional neural network fault diagnosis models are compared, and it is concluded that the proposed method is better in terms of fault diagnosis accuracy and stability.

Soft Voting-based Ensemble Model for Bengali Sign Gesture Recognition

Human hand gestures are becoming one of the most important, intuitive, and essential means of recognizing sign language. Sign language is used to convey different meanings through visual-manual methods. Hand gestures help the hearing impaired to communicate. Nevertheless, it is very difficult to achieve a high recognition rate of hand gestures due to the environment and physical anatomy of human beings such as light condition, hand size, position, and uncontrolled environment. Moreover, the recognition of appropriate gestures is currently considered a major challenge. In this context, this paper proposes a probabilistic soft voting-based ensemble model to recognize Bengali sign gestures. We have divided this study into pre-processing, data augmentation and ensemble model-based voting process, and classification for gesture recognition. The purpose of pre-processing is to remove noise from input images, resize it, and segment hand gestures. Data augmentation is applied to create a larger database for in-depth model training. Finally, the ensemble model consists of a support vector machine (SVM), random forest (RF), and convolution neural network (CNN) is used to train and classify gestures. Whereas, the ReLu activation function is used in CNN to solve neuron death problems and to accelerate RF classification through principal component analysis (PCA). A Bengali Sign Number Dataset named “BSN-Dataset” is proposed for model performance. The proposed technique enhances sign gesture recognition capabilities by utilizing segmentation, augmentation, and soft-voting classifiers which have obtained an average of 99.50% greater performance than CNN, RF, and SVM individually, as well as significantly more accuracy than existing systems.

Direct data-driven forecast of local turbulent heat flux in Rayleigh–Bénard convection

A combined convolutional autoencoder–recurrent neural network machine learning model is presented to directly analyze and forecast the dynamics and low-order statistics of the local convective heat flux field in a two-dimensional turbulent Rayleigh–Bénard convection flow at Prandtl number Pr=7 and Rayleigh number Ra=107. Two recurrent neural networks are applied for the temporal advancement of turbulent heat transfer data in the reduced latent data space, an echo state network, and a recurrent gated unit. Thereby, our work exploits the modular combination of three different machine learning algorithms to build a fully data-driven and reduced model for the dynamics of the turbulent heat transfer in a complex thermally driven flow. The convolutional autoencoder with 12 hidden layers is able to reduce the dimensionality of the turbulence data to about 0.2% of their original size. Our results indicate a fairly good accuracy in the first- and second-order statistics of the convective heat flux. The algorithm is also able to reproduce the intermittent plume-mixing dynamics at the upper edges of the thermal boundary layers with some deviations. The same holds for the probability density function of the local convective heat flux with differences in the far tails. Furthermore, we demonstrate the noise resilience of the framework. This suggests that the present model might be applicable as a reduced dynamical model that delivers transport fluxes and their variations to coarse grids of larger-scale computational models, such as global circulation models for atmosphere and ocean.

Power Generation Prediction of Building-Integrated Photovoltaic System with Colored Modules Using Machine Learning

The building-integrated photovoltaic (BIPV) system is provoking mention as a technology for generating the energy consumed in cities with renewable sources. As the number of BIPV systems increases, performance diagnosis through power-generation predictions becomes more essential. In the case of a colored BIPV module that has been installed on a wall, it is more difficult to predict the amount of power generation because the shading loss varies based on the entrance altitude of the irradiance. Recently, artificial intelligence technology that is able to predict power by learning the output data of the system has begun being used. In this paper, the power values of colored BIPV systems that have been installed on walls are predicted, and the system output values are compared. The current-voltage (I–V) curve data are measured to predict the power required changing the intensity of the irradiance, and the linear regression model is derived for the changes in the voltage and current at a maximum power operating point and during irradiance changes. To improve the power prediction accuracy by considering the shading loss of colored BIPVs, a new model is proposed via neural network machine learning (ML). In addition, the accuracy of the proposed prediction models is evaluated by comparing the metrics such as RMSE, MAE, and R2. As a result of testing the linear regression model and the proposed ML model, the R2 values for the voltage and current values of the proposed ML model were 5% higher for voltage and 2% higher for current. From this result, the proposed ML model of the RMSE about real power improved by more than 50% (0.0754 kW) compared to the simulation model (0.1581 KW). The proposed model demonstrates high-accuracy power estimations and is expected to help diagnose the performance of BIPV systems with colored modules.

A Data-Driven Fault Diagnosis Method for Solid Oxide Fuel Cell Systems

In this study, a data-driven fault diagnosis method was developed for solid oxide fuel cell (SOFC) systems. First, the complete experimental data was obtained following the design of the SOFC system experiments. Then, principal component analysis (PCA) was performed to reduce the dimensionality of the obtained experimental data. Finally, the fault diagnosis algorithms were designed by support vector machine (SVM) and BP neural network to identify and prevent the reformer carbon deposition and heat exchanger rupture faults, respectively. The research results show that both SVM and BP fault diagnosis algorithms can achieve online fault identification. The PCA + SVM algorithm was compared with the SVM algorithm, BP algorithm, and PCA + BP algorithm, and the results show that the PCA + SVM algorithm is superior in terms of running time and accuracy, the diagnosis accuracy reached more than 99%, and the running time was within 20 s. The corresponding system optimization scheme is also proposed.

Prediction of Hospital Intesive Patients Using Neural Network Algorithm

This study aims to predict whether the patient deserves to be inpatient or outpatient by comparing several machine learning techniques, namely, logistic regression, decision tree, neural network, random forest, gradient boosting. The research method uses three stages of research, namely data collection, data preprocessing, and data modeling. Implementation of program code using google colab and python programming language. The dataset used as the research sample is Electronic Health Record Predicting data. Based on the accuracy results generated in this study, the use of the Neural Network machine learning algorithm to predict hospitalization decisions for patients has proven to be a machine learning algorithm that has the highest accuracy rate reaching 74, 47% compared to other comparison machine learning algorithms, namely logistic regression, decision tree, neural network, random forest, gradient boosting.

Research on enterprise financial economics early warning based on machine learning method

The stability of the economic market is an important factor for the rapid development of the economy, especially for the listed companies, whose financial and economic stability affects the stability of the financial market. It is helpful for the healthy development of enterprises and financial markets to make an accurate early warning of the financial economy of listed enterprises. This paper briefly introduced the support vector machine (SVM) and back-propagation neural network (BPNN) algorithms in the machine learning method. To make up for the defects of the two algorithms, they were combined and applied to the enterprise financial economics early warning. A simulation experiment was carried out on the single SVM algorithm-based, single BPNN algorithm-based, and SVM algorithm and BPNN algorithm combined model with the MATLAB software. The results show that the SVM algorithm and BP algorithm combined model converges faster and has higher precision and recall rate and larger area under the curve (AUC) than the single SVM algorithm-based model and the single BPNN algorithm-based model.

Estimation of bulk electrical conductivity in saline medium with contaminated lead solution through TDR coupled with machine learning

Time-domain reflectometry (TDR) has been used for the characterization of media; however, the results of TDR tests significantly differ according to the types of solutions. The objective of this study is to suggest a new relationship between TDR output values and bulk electrical conductivity based on a machine learning algorithm for enhancing the reliability of TDR measurement. Various salinities (0%, 1%, 2%, and 3%) and lead concentrations (0, 0.5, 1, 2, 5, and 10 mg/L) are applied along with silica sand, classified as SP in USCS, to create media. A laboratory test is performed to measure the TDR waveform at the bottom of the cylindrical cell, and a resistance probe is also installed to obtain the true bulk electrical conductivity in the cell. A deep neural network machine learning algorithm is applied to establish the relationship between the TDR output value and the bulk electrical conductivity at each frequency of 0.1, 0.12, 1, 10, and 100 kHz. The highly important variables are also defined through random forest. This study demonstrates that the TDR can be reliably converted into bulk electrical conductivity when two different solutions are mixed.

Reliability analysis with corrosion defects in submarine pipeline case study: Oil pipeline in Ab-khark island

The offshore industry widely uses offshore pipelines to transport petroleum products due to their high capacity and reasonable installation costs. Corrosion is the leading failure cause of offshore pipelines, and their reliability analysis is necessary for reducing maintenance costs. This study presents a corrosion analysis in the bursting mode of the pipeline using a codified method based on probabilistic space theory, considering the monitored data of the oil pipeline and the pigging data of the AB-KHARK pipeline. The relationship between pigging data, parameters affecting corrosion, and the reduction of pipe wall thickness are determined based on the indicators of Regression and Neural Network machine learning methods. A new method is presented for modeling the failure modes of offshore pipelines based on their historical data and statistical computation results. Finally, the system is investigated by the first-order reliability method considering the limit state function provided by evaluating the offshore pipeline reliability. The proposed method is evaluated using a dataset from AB-KHARK island in IRAN indicate zero percentage error in failure prediction versus related works with 1% error for PRCI and MB methods, 6, 8, and 11 percentages error against RITCHE, DNV, and AMSE methods, respectively.

Prediction of Early Compressive Strength of Ultrahigh-Performance Concrete Using Machine Learning Methods

In this study, a new prediction model is proposed to predict the 7-day compressive strength of ultrahigh-performance concrete (UHPC) with different mix proportions using artificial neural network (ANN) and support vector machine (SVM). The predicted results are compared with the experimental results to verify the proposed model. Then, the importance of each component and the sensitivity of parameters are investigated. The research proves that the proposed model can estimate the 7-day compressive strength of UHPC based on the mix proportions.

Detection of Electrical Fault in Medium Voltage Installation Using Support Vector Machine and Artificial Neural Network

Detection of Electrical Fault in Medium Voltage Installation Using Support Vector Machine and Artificial Neural Network

Machine learning for percolation utilizing auxiliary Ising variables.

Machine learning for phase transition has received intensive research interest in recent years. However, its application in percolation still remains challenging. We propose an auxiliary Ising mapping method for the machine learning study of the standard percolation as well as a variety of statistical mechanical systems in correlated percolation representation. We demonstrate that unsupervised machine learning is able to accurately locate the percolation threshold, independent of the spatial dimension of system or the type of phase transition, which can be first-order or continuous. Moreover, we show that, by neural network machine learning, auxiliary Ising configurations for different universalities can be classified with a high confidence level. Our results indicate that the auxiliary Ising mapping method, despite its simplicity, can advance the application of machine learning in statistical and condensed-matter physics.

Automatic detection of passing and shooting in water polo using machine learning: a feasibility study

ABSTRACT There is currently no efficient way to quantify overhead throwing volume in water polo. Therefore, this study aimed to test the feasibility of a method to detect passes and shots in water polo automatically using inertial measurement units (IMU) and machine-learning algorithms. Eight water polo players wore one IMU sensor on the wrist (dominant hand) and one on the sacrum during six practices each. Sessions were filmed with a video camera and manually tagged for individual shots or passes. Data were synchronised between video tagging and IMU sensors using a cross-correlation approach. Support vector machine (SVM) and artificial neural networks (ANN) were compared based on sensitivity and specificity for identifying shots and passes. A total of 7294 actions were identified during the training sessions, including 945 shots and 5361 passes. Using SVM, passes and shots together were identified with 94.4% (95%CI = 91.8–96.4) sensitivity and 93.6% (95%CI = 91.4–95.4) specificity. Using ANN yielded similar sensitivity (93.0% [95%CI = 90.1–95.1]) and specificity (93.4% [95%CI = 91.1 = 95.2]). The results suggest that this method of identifying overhead throwing motions with IMU has potential for future field applications. A set-up with one single sensor at the wrist can suffice to measure these activities in water polo.

Classifying Conduct Disorder Using a Biopsychosocial Model and Machine Learning Method

BackgroundConduct disorder (CD) is a common syndrome with far-reaching effects. Risk factors for the development of CD span social, psychological, and biological domains. Researchers note that predictive models of CD are limited if the focus is on a single risk factor or even a single domain. Machine learning methods are optimized for the extraction of trends across multidomain data but have yet to be implemented in predicting the development of CD. MethodsSocial (e.g., family, income), psychological (e.g., psychiatric, neuropsychological), and biological (e.g., resting-state graph metrics) risk factors were measured using data from the baseline visit of the Adolescent Brain Cognitive Development Study when youth were 9 to 10 years old (N = 2368). Applying a feed-forward neural network machine learning method, risk factors were used to predict CD diagnoses 2 years later. ResultsA model with factors that included social, psychological, and biological domains outperformed models representing factors within any single domain, predicting the presence of a CD diagnosis with 91.18% accuracy. Within each domain, certain factors stood out in terms of their relationship to CD (social: lower parental monitoring, more aggression in the household, lower income; psychological: greater attention-deficit/hyperactivity disorder and oppositional defiant disorder symptoms, worse crystallized cognition and card sorting performance; biological: disruptions in the topology of subcortical and frontoparietal networks). ConclusionsThe development of an accurate, sensitive, and specific predictive model of CD has the potential to aid in prevention and intervention efforts. Key risk factors for CD appear best characterized as reflecting unpredictable, impulsive, deprived, and emotional external and internal contexts.

Classification of Iris Flower by Random Forest Algorithm

With the introduction of artificial intelligence into our lives, artificial intelligence researches and applications in different fields such as agriculture, health, military and engineering applications have become very popular iris flower was classified using the popular Random Forest, support vector machine and Artificial neural network machine learning classifiers with high accuracy rates. As a result of the classification, the performance of the trained models was evaluated according to the confusion matrix, sensitivity, specificity, accuracy, F1 score, ROC curve and AUC evaluation criteria. The random forest algorithm was the most successful among the trained algorithms with an accuracy rate of 97%.