A support vector machine integrated system for the classification of operation anomalies in nuclear components and systems

Prediction of steam/water stratified flow characteristics in NPPs transients using SVM learning algorithm with combination of thermal-hydraulic model and new data mapping technique

본 논문은 조명의 변화가 심한 영상에서 손 형태를 안정적으로 인지하는 기법에 관한 것이다. 제안한 방법은 HSI 색상공간에서 색상(Hue) 및 색상 기울기(Hue-Gradient)를 기반으로 정의된 배경모델을 구축하고, 실시간으로 입력되는 영상과의 배경차분(background subtraction)기법을 이용하여 배경과 손을 구분한다. 추출된 손의 영역으로부터 18가지의 특징요소를 추출하고 이를 기반으로 다중클래스 SVM(Support Vector Machine) 학습 기법을 사용하여 손의 형태를 인지한다. 제안 기법은 색상 기울기를 배경 차분에 적용함으로써, 조명 환경이 배경 모델의 조명과 다르게 급격한 변화가 이루어졌을 때도 안정적으로 손의 윤곽정보를 추출할 수 있도록 하였다. 또한, 실시간 처리를 저해하는 복잡한 손의 특성정보 대신, OBB의 크기에 대하여 정규화된 두 개의 고유값과 객체 기반 바운딩 박스(OBB)를 구성하는 16개 세부 영역에서의 손 윤곽픽셀의 개수를 손의 특성정보로 사용하였다. 본 논문에서는 급격한 조명 변화 상황에서 기존 RGB 색상요소를 기반으로 하는 배경차분법과 색상을 기반으로 하는 배경차분법, 본 논문에서 제안하는 색상 기울기 기반 배경 차분법의 결과를 비교함으로써 제안 기법의 안정성을 입증하였다. 6명의 실험대상자의 1부터 9까지의 수지화 2700개의 영상으로부터 손 특성 정보를 추출하고 이에 대하여 훈련을 통한 학습 모델을 생성하였다. 학습모델을 기반으로 실험자 6인의 손 형태 1620개의 데이터에 대하여 인지 실험을 실시하여 92.6%에 이르는 손 형태 인식 성공률을 얻었다. In this paper, we present a robust hand recognition approach to sudden illumination changes. The proposed approach constructs a background model with respect to hue and hue gradient in HSI color space and extracts a foreground hand region from an input image using the background subtraction method. Eighteen features are defined for a hand pose and multi-class SVM(Support Vector Machine) approach is applied to learn and classify hand poses based on eighteen features. The proposed approach robustly extracts the contour of a hand with variations in illumination by applying the hue gradient into the background subtraction. A hand pose is defined by two Eigen values which are normalized by the size of OBB(Object-Oriented Bounding Box), and sixteen feature values which represent the number of hand contour points included in each subrange of OBB. We compared the RGB-based background subtraction, hue-based background subtraction and the proposed approach with sudden illumination changes and proved the robustness of the proposed approach. In the experiment, we built a hand pose training model from 2,700 sample hand images of six subjects which represent nine numerical numbers from one to nine. Our implementation result shows 92.6% of successful recognition rate for 1,620 hand images with various lighting condition using the training model.

The paper explores the potential of Support Vector Machines (SVM) approach in predicting 28-day compressive strength and slump flow of self-compacting concrete. Total of 80 data collected from the exiting literature were used in present work. To compare the performance of the technique, prediction was also done using a back propagation neural network model. For this data-set, RBF kernel worked well in comparison to polynomial kernel based support vector machines and provide a root mean square error of 4.688 (MPa) (correlation coefficient=0.942) for 28-day compressive strength prediction and a root mean square error of 7.825 cm (correlation coefficient=0.931) for slump flow. Results obtained for RMSE and correlation coefficient suggested a comparable performance by Support Vector Machine approach to neural network approach for both 28-day compressive strength and slump flow prediction.

Educational data mining provides the process of applying different data mining tools and techniques to analyze and visualize the data of an institution and can be used to discover a unique pattern of students' academic performance. Secondary schools are increasing rapidly from past years and progress of the institution can be measured based on student's success and failure rates. Failure rates can be measured in terms of a core subject such as mathematics which has been considered in this proposed system. Real data was collected using school reports and questionnaire method by the Portugal school which has been used for the project. The students will be classified according to the grades assigned for the range of marks scored by them. This involves classifying the students into five levels of grading system starting from grade ‘A’ to grade ‘F’ where grade ‘A’ represents the student getting the highest marks, grade ‘B’ being the second highest, grade ‘C’ being third, grade ‘D’ being the fourth and the grade ‘F’ which implies that the student has failed. To carry out this type of classification many machine learning algorithms can be used to implement it. Comparison was made between the algorithms like Multiclass Support Vector Machine and Neural Networks using the Weka tool. Based on the analysis carried out Multiclass Support Vector Machine showed prominent accuracy. The Multiclass Support Vector Machine is implemented on the basis of one-to-rest strategy use of class labels which is primarily an extension of linear Support Vector Machine. In order to get appropriate results in terms of the accuracy of the model, parameters like ‘C’ and ‘Gamma’ is tuned while implementing Multiclass Support Vector Machine. The result gives a good predictive accuracy based on the grades that are provided by the school. The accuracy of prediction made by Support Vector Machine Classifier is determined by using K-fold cross validation, according to which the accuracy is 89%.

Software project development has high failure rate. Software project risk management may gain a high rate of return in investment. Establishing an intelligent risk evaluation model for project will be valuable in the analysis and control of project risks. In this paper, we employed neural network (NN) and support vector machine (SVM) approaches to establish a model for risk evaluation in project development. In the model, the input is a vector of software risk factors that were obtained through interview with 30 experts, and the output is the final outcome of the project. The data for modeling were collected from 120 real software projects through questionnaires. The experiment shows the model is valid. Interestingly, SVM is a powerful supervised learning method, and some believe that it is a more promising classification method that may someday supercede NN. In our study, the standard neural network model had lower prediction accuracy compared to SVM due to its tendency in finding local optima. However, after attempt in optimizing the neural network model with genetic algorithm, the experimental results showed that our enhanced model surpassed SVM in performance.

Venn Predictors (VPs) are machine learning algorithms that can provide well calibrated multiprobability outputs for their predictions. The only drawback of Venn Predictors is their computational inefficiency, especially in the case of large datasets. In this work, we propose an Inductive Venn Predictor (IVP) which overcomes the computational inefficiency problem of the original Venn Prediction framework. Each VP is defined by a taxonomy which separates the data into categories. We develop an IVP with a taxonomy derived from a multiclass Support Vector Machine (SVM), and we compare our method with other probabilistic methods for SVMs, namely Platt’s method, SVM Binning, and SVM with Isotonic Regression. We show that these methods do not always provide well calibrated outputs, while our IVP will always guarantee this property under the i.i.d. assumption.KeywordsSupport Vector Machinewell calibrated probabilitiesmulticlassInductive Venn PredictorMachine Learning

Nuclear systems and reactors stand out by their intrinsic complexity and the wide range of scientific and technical disciplines they involve. The design, modeling, and operation of a reactor rely on the close interdependence between neutronics, thermal-hydraulics, and thermomechanics, complemented by the description of the thermodynamic cycle of the secondary circuit and the plant's control and instrumentation systems. Historically, each of these fields has led to the development of specialized computational codes, recognized for their reliability and accuracy. However, integrating these tools into comprehensive simulation frameworks has quickly raised major challenges related to interoperability, data consistency, and the heterogeneity of physical and numerical models. To address these challenges, the standardization of code coupling interfaces has attracted growing interest, particularly following the release of the Functional Mock-up Interface (FMI) standard. This standard provides a unified framework for the systematic coupling of models from different domains, thereby facilitating communication between heterogeneous software tools. In this context, three coupling strategies were developed, analyzed, and applied to various nuclear systems: coupling through a heat exchanger between two thermal systems, pressure-velocity coupling between distinct thermal-hydraulic circuits, and a multiscale coupling linking the reactor core to detailed fuel pin models. These coupling strategies were illustrated through several representative case studies. The first application focused on the simulation of load-follow operation in a simplified nuclear power plant. The second addressed the modeling of a simplified nuclear propulsion system during a trans-lunar injection. Finally, the third examined the effects of beam trips on the fuel pins of an Accelerator Driven System. Overall, these studies demonstrate the feasibility and relevance of multi-code coupling based on the FMI standard for the integrated simulation of complex nuclear systems. The results confirm that this approach represents a promising pathway toward more flexible and interoperable modeling tools for reactor design and safety analysis.

This paper will study insurance data clustering using Support Vector Machine (SVM) approaches. It investigates the optimum condition employing the three most popular kernels of SVM, i.e., linear, polynomial, and radial basis kernel. To explore sum insured datasets, kernel comparisons for Root Mean Square Error (RMSE) and density analysis have been provided. It employs these kernels to classify based on sum insured datasets. The objective of this research is to demonstrate to industrial researchers that data grouping may be accomplished in an organized, error-free, and efficient manner utilizing R programming and the SVM approach. In this study, we check the insurance data for the sum insured with statistical methods in the form of Model Performance Evaluation (MPE), Receiver Operating Characteristics (ROC), Area Under Curve (AUC), partial AUC (pAUC), smoothing, confidence intervals, and thresholds. Then, sum insured data are followed up to classify using SVM kernels. This paper finds new ideas for evaluating insurance data using the SVM approach with multiple kernels. This novel research emphasizes the statistical analysis methods for insurance data and uses the SVM method for more accurate data classification. Finally, it informs that this research is a pure finding, and there has never been any research on this subject. This research was conducted using the sum insured data as a sample from the Office of the Insurance Commission (OIC) in Thailand as an independent insurance institution providing actual data. Doi: 10.28991/ESJ-2022-06-04-014 Full Text: PDF

A support vector machine (SVM) approach was presented for predicting the drilling fluid density at high temperature and high pressure (HTHP). It is a universal model for water-based, oil-based, and synthetic drilling fluids. Available experimental data in the literature were used to develop and test this SVM model. Good agreement between SVM predictions and measured drilling fluid density values confirmed that the developed SVM model had good predictive precision and extrapolative features. The SVM model was also compared with the most popular models such as the artificial neural network (ANN) model, empirical correlations, and analytical models. Results showed that the SVM approach outperformed the competing methods for the prediction of drilling fluid density at HTHP.

In this paper, we present an evaluation of face recognition algorithm in video surveillance system using multi-class Support Vector Machine (SVM). We also show how the variation of sizes (due to distance) and intensity of training images (due to camera hardware) can improve the face recognition rate. The system starts by detecting the face regions using AdaBoost training algorithm. Holistic approach is chosen to extract features from the detected face regions as this approach is the simplest. After the feature vectors are generated, multi-class SVM is used for classification step. SVM is chosen as the classifier as it is a low samples, high dimensional classifier. As the training and test set, we use SCface database in which the images were taken in uncontrolled indoor environment using five video surveillance cameras of various qualities. Based on our study using only linear kernel multi-class SVM, it is shown that distance of approximately 2.6m gives the best results, and the training and test images have to be from similar intensity cameras although surveillance camera will have different quality of images.

Recently, focus on maintenance strategies has been shifted towards prognostic health management (PHM) and a number of state of the art algorithms based on data-driven prognostics have been developed to predict the health states of degrading components based on sensory data. Amongst these algorithms, Multiclass Support Vector Machines (MC-SVM) has gained popularity due to its relatively high classification accuracy, ability to classify multiple patterns and capability to handle noisy / incomplete data. However, its application is limited by the difficulty in determining the required kernel function and penalty parameters. To address this problem, this paper proposes a hybrid differential evolution – particle swarm optimization (DE-PSO) algorithm to optimize the MC-SVM kernel function and penalty parameters. The differential algorithm (DE) obtains the search limit for the SVM parameters, while the particle swarm optimization algorithm (PSO) determines the global optimum parameters for a given training data set. Since degrading machinery components display several degradation stages in their lifetime, the MC-SVM trained with optimum parameters are used to estimate the health states of a degrading machinery component, from which the remaining useful life (RUL) is predicted. This method improves the classification accuracy of MC-SVM in predicting the health states of a machinery component and consequently increases the accuracy of RUL predictions. The feasibility of the method is validated using bearing prognostic run-to-failure data obtained from NASA public data repository. A comparative study between MC-SVM with parameters obtained using simple grid search with n-fold cross validation and MC- SVM with DE-PSO based on prognostic performance metrics reveals that the proposed method has better performance, with all the cases considered falling within a 10 % error margin. The method also outperforms other soft computing methods proposed in literature.

Undernutrition among children under the age of five is a major public health concern, especially in developing countries. This study aimed to use machine learning (ML) algorithms to predict undernutrition and identify its associated factors. Secondary data analysis of the 2017 Multiple Indicator Cluster Survey (MICS) was performed using R and Python. The main outcomes of interest were undernutrition (stunting: height-for-age (HAZ) < -2 SD; wasting: weight-for-height (WHZ) < -2 SD; and underweight: weight-for-age (WAZ) < -2 SD). Seven ML algorithms were trained and tested: linear discriminant analysis (LDA), logistic model, support vector machine (SVM), random forest (RF), least absolute shrinkage and selection operator (LASSO), ridge regression, and extreme gradient boosting (XGBoost). The ML models were evaluated using the accuracy, confusion matrix, and area under the curve (AUC) receiver operating characteristics (ROC). In total, 8564 children were included in the final analysis. The average age of the children was 926 days, and the majority were females. The weighted prevalence rates of stunting, wasting, and underweight were 17%, 7%, and 12%, respectively. The accuracies of all the ML models for wasting were (LDA: 84%; Logistic: 95%; SVM: 92%; RF: 94%; LASSO: 96%; Ridge: 84%, XGBoost: 98%), stunting (LDA: 86%; Logistic: 86%; SVM: 98%; RF: 88%; LASSO: 86%; Ridge: 86%, XGBoost: 98%), and for underweight were (LDA: 90%; Logistic: 92%; SVM: 98%; RF: 89%; LASSO: 92%; Ridge: 88%, XGBoost: 98%). The AUC values of the wasting models were (LDA: 99%; Logistic: 100%; SVM: 72%; RF: 94%; LASSO: 99%; Ridge: 59%, XGBoost: 100%), for stunting were (LDA: 89%; Logistic: 90%; SVM: 100%; RF: 92%; LASSO: 90%; Ridge: 89%, XGBoost: 100%), and for underweight were (LDA: 95%; Logistic: 96%; SVM: 100%; RF: 94%; LASSO: 96%; Ridge: 82%, XGBoost: 82%). Age, weight, length/height, sex, region of residence and ethnicity were important predictors of wasting, stunting and underweight. The XGBoost model was the best model for predicting wasting, stunting, and underweight. The findings showed that different ML algorithms could be useful for predicting undernutrition and identifying important predictors for targeted interventions among children under five years in Ghana.

Support Vector Machine (SVM) is mainly used to classify the data into two categories. To solve the multi-category problems using SVM, researchers used two approaches. The first approach based on solving multiple SVM binary classifiers, whereas another approach based on solving a single optimization problem. In this paper, we have used the first approach and proposed an Efficient Multiclass Support Vector Machine (ESVM) algorithm using a skewed binary tree. To construct the skewed binary tree, no extra efforts are required as compared to the binary tree approach. The algorithm is tested on the benchmark data sets, and the results are compared with both the multiclass approaches of SVM. The ESVM’s results are compared with five techniques of solving multiple binary SVM classifiers and four techniques of solving a single optimization problem. The comparative experiments prove the efficiency of the ESVM in terms of its accuracy as compared to other contemporary algorithms. Further, ESVM is successfully applied for classification of the email dataset into positive, negative and neutral sentiments.

The problem of conditional simulation of non-Gaussian stochastic processes and fields has gained a significant interest recently because of its applications in many fields, such as wind engineering, ocean engineering, and soil engineering. In this paper, the support vector machines (SVM) approach is developed for the conditional simulation of non-Gaussian stochastic processes and fields. To show the advantages of the presented method, the conditional simulation of non-Gaussian fluctuating wind pressures is carried out by using SVM and artificial neural networks (ANN). SVM considers three kinds of kernel function, such as linear function, Gaussian radial basis function, and exponential radial basis function, whereas ANN employs back-propagation and generalized regression. In machine learning of these artificial intelligences, two ways (interpolation and extrapolation) are employed to train finite non-Gaussian samples. The feasibility and validity of these algorithms are evaluated through the correlation coefficients, root mean square errors, skewness errors, and kurtosis errors between simulated samples and target samples and probability density functions (PDF), power spectral density (PSD) functions, and autocorrelation functions (ACF) of the simulated non-Gaussian fluctuating wind pressures versus their corresponding targets. The results show that the accuracy and effectiveness of SVM with an appropriate kernel function are superior to the back-propagation neural network (BPNN) and generalized regression neural network (GRNN). Furthermore, the advantage of the presented SVM approach is very obvious when the trained non-Gaussian samples are few.

Bornean orangutans are critically endangered, and experts are working to manually spot and classify their nests. This is a challenging task, as it requires a deep understanding of orangutan behavior and their natural habitats. To address this challenge, researchers are developing advanced image analysis techniques that can minimize human intervention. In this paper, we propose a method for classifying orangutan nests using Support Vector Machines (SVMs). SVMs are a powerful machine learning algorithm that has been shown to be effective for image classification. Our method includes a comprehensive image pre-processing and enhancement pipeline, as well as a rigorous training and testing procedure. The results of our study show that SVMs can achieve high accuracy in classifying orangutan nests. We achieved an accuracy of 79.90%, a F1-score of 79.87%, a precision of 79.91%, and a recall of 79.82%. These results demonstrate the effectiveness of our method and the potential of SVMs for this task.