Support Vector Machine Framework Research Articles

Support Vector Machine Framework for Detecting Ambiguous Contours of Space Debris Cloud

Operational spacecraft are prepared for collisions with space debris by installing protective walls (bumpers). A group of vaporized and liquefied microparticles generated by debris collisions is called a debris cloud. To evaluate the protective performance of bumpers, it is important to measure the spread velocity and angle of the debris cloud. However, debris cloud images exhibit ambiguous contours of debris clouds. Consequently, measurement errors in the spread velocity and spread angles occur because of the observer’s subjectivity. Commonly used derivative-based methods can detect steep contours; however, they cannot detect ambiguous contours such as those of debris clouds. In this study, a method for uniquely detecting the ambiguous contours of a debris cloud based on supervised machine learning and a continuous wavelet transform (CWT) is proposed. Feature points that indicate candidate points of the debris cloud contour are extracted using a method based on CWT. Four types of specific features for debris cloud contours were assigned to the feature points. Supervised machine learning using a support vector machine can detect the debris cloud contours. The proposed method enables the evaluation of the ambiguous contours of a debris cloud and the protection performance of a bumper without observer subjectivity.

Heart Sound Processing for Early Diagnostic of Heart Abnormalities using Support Vector Machine

This paper addresses the critical issue of cardiovascular disease (CVD), the leading cause of global mortality, emphasizing the imperative for effective and early detection to mitigate CVD-related deaths. The research problem underscores the urgency of developing advanced diagnostic tools to identify heart abnormalities promptly. The primary objective is to create a Support Vector Machine (SVM) algorithm for accurate classification of different heart conditions, namely Normal heart, Mitral Stenosis, and Mitral Regurgitation. To achieve this objective, the study utilizes a dataset of heart sounds available online using a 10-fold cross-validation method. The focus is on evaluating the efficacy of various kernel functions within the SVM framework for heart sound classification. The findings demonstrate that the linear kernel exhibits superior accuracy and robustness in effectively classifying heart conditions. Notably, the proposed classification method attains an impressive 96% accuracy, highlighting its potential as a reliable tool for early detection of cardiovascular diseases. This research contributes to the ongoing efforts to enhance diagnostic capabilities and ultimately reduce the global burden of CVD-related fatalities.

A novel bounded loss framework for support vector machines

This paper introduces a novel bounded loss framework for SVM and SVR. Specifically, using the Pinball loss as an illustration, we devise a novel bounded exponential quantile loss (Leq-loss) for both support vector machine classification and regression tasks. For Leq-loss, it not only enhances the robustness of SVM and SVR against outliers but also improves the robustness of SVM to resampling from a different perspective. Furthermore, EQSVM and EQSVR were constructed based on Leq-loss, and the influence functions and breakdown point lower bounds of their estimators are derived. It is proved that the influence functions are bounded, and the breakdown point lower bounds can reach the highest asymptotic breakdown point of 1/2. Additionally, we demonstrated the robustness of EQSVM to resampling and derived its generalization error bound based on Rademacher complexity. Due to the Leq-loss being non-convex, we can use the concave–convex procedure (CCCP) technique to transform the problem into a series of convex optimization problems and use the ClipDCD algorithm to solve these convex optimization problems. Numerous experiments have been conducted to confirm the effectiveness of the proposed EQSVM and EQSVR.

Intelligent optimization of steam gasification catalysts for palm oil waste using support vector machine and adaptive transition marine predator algorithm

The present study investigates the optimization of the steam gasification process for the conversion of palm oil waste into environmentally friendly energy, utilizing the catalytic properties of calcium oxide and coal bottom ash. The objective of our research is to investigate the enhancement of the conversion process by employing a machine-learning approach. Specifically, we utilize a support vector machine (SVM) to model and evaluate the impact of different operational parameters on the resulting gas mixture. One notable feature of this study involves the incorporation of an adaptive marine predator algorithm (AMPA) into the SVM framework, aiming to enhance the predicted precision and efficiency of the model. The primary focus of this study revolves around the development of an intelligent optimization framework that surpasses conventional machine learning techniques, hence providing a more dynamic and efficient strategy for process improvement. The SVM model’s performance, as assessed against experimental benchmarks, exhibits a notable degree of predictive accuracy and substantial concurrence with observed data. This increase in performance indicates that our methodology has the potential to make a significant contribution to the enhancement of renewable catalysts in gasification processes. The findings of this study could potentially have significant ramifications for the advancement of renewable energy production and the creation of intelligent systems in complicated industrial applications.

BAFL-SVM: A blockchain-assisted federated learning-driven SVM framework for smart agriculture

The combination of blockchain and Internet of Things technology has made significant progress in smart agriculture, which provides substantial support for data sharing and data privacy protection. Nevertheless, achieving efficient interactivity and privacy protection of agricultural data remains a crucial issues. To address the above problems, we propose a blockchain-assisted federated learning-driven support vector machine (BAFL-SVM) framework to realize efficient data sharing and privacy protection. The BAFL-SVM is composed of the FedSVM-RiceCare module and the FedPrivChain module. Specifically, in FedSVM-RiceCare, we utilize federated learning and SVM to train the model, improving the accuracy of the experiment. Then, in FedPrivChain, we adopt homomorphic encryption and a secret-sharing scheme to encrypt the local model parameters and upload them. Finally, we conduct a large number of experiments on a real-world dataset of rice pests and diseases, and the experimental results show that our framework not only guarantees the secure sharing of data but also achieves a higher recognition accuracy compared with other schemes.

Interval-based KKT framework for support vector machines and beyond

Our article proves inequalities for interval optimization and shows that feasible and descent directions do not intersect in constrained cases. Mainly, we establish some new interval inequalities for interval-valued functions by defining LC-partial order. We use LC-partial order to study Karush–Kuhn–Tucker (KKT) conditions and expands Gordan's theorems for interval linear inequality systems. By applying Gordan's theorem, we can determine the best outcomes for interval optimization problems (IOPs) that have constraints, such as Fritz John and KKT conditions. The optimality conditions are observed with inclusion relations rather than equality. We can use the KKT condition for binary classification with interval data and support vector machines(SVMs). We present some examples to illustrate our results.

Semi-supervised learning with constrained virtual support vector machines for classification of remote sensing image data

We introduce two semi-supervised models for the classification of remote sensing image data. The models are built upon the framework of Virtual Support Vector Machines (VSVM). Generally, VSVM follow a two-step learning procedure: A Support Vector Machines (SVM) model is learned to determine and extract labeled samples that constitute the decision boundary with the maximum margin between thematic classes, i.e., the Support Vectors (SVs). The SVs govern the creation of so-called virtual samples. This is done by modifying, i.e., perturbing, the image features to which a decision boundary needs to be invariant. Subsequently, the classification model is learned for a second time by using the newly created virtual samples in addition to the SVs to eventually find a new optimal decision boundary. Here, we extend this concept by (i) integrating a constrained set of semi-labeled samples when establishing the final model. Thereby, the model constrainment, i.e., the selection mechanism for including solely informative semi-labeled samples, is built upon a self-learning procedure composed of two active learning heuristics. Additionally, (ii) we consecutively deploy semi-labeled samples for the creation of semi-labeled virtual samples by modifying the image features of semi-labeled samples that have become semi-labeled SVs after an initial model run. We present experimental results from classifying two multispectral data sets with a sub-meter geometric resolution. The proposed semi-supervised VSVM models exhibit the most favorable performance compared to related SVM and VSVM-based approaches, as well as (semi-)supervised CNNs, in situations with a very limited amount of available prior knowledge, i.e., labeled samples.

Robust Twin Bounded Support Vector Classifier With Manifold Regularization.

Support vector machine (SVM), as a supervised learning method, has different kinds of varieties with significant performance. In recent years, more research focused on nonparallel SVM, where twin SVM (TWSVM) is the typical one. In order to reduce the influence of outliers, more robust distance measurements are considered in these methods, but the discriminability of the models is neglected. In this article, we propose robust manifold twin bounded SVM (RMTBSVM), which considers both robustness and discriminability. Specifically, a novel norm, that is, capped L₁-norm, is used as the distance metric for robustness, and a robust manifold regularization is added to further improve the robustness and classification performance. In addition, we also use the kernel method to extend the proposed RMTBSVM for nonlinear classification. We introduce the optimization problems of the proposed model. Subsequently, effective algorithms for both linear and nonlinear cases are proposed and proved to be convergent. Moreover, the experiments are conducted to verify the effectiveness of our model. Compared with other methods under the SVM framework, the proposed RMTBSVM shows better classification accuracy and robustness.

Sparse optimization via vector k-norm and DC programming with an application to feature selection for support vector machines

Sparse optimization is about finding minimizers of functions characterized by a number of nonzero components as small as possible, such paradigm being of great practical relevance in Machine Learning, particularly in classification approaches based on support vector machines. By exploiting some properties of the k-norm of a vector, namely, of the sum of its k largest absolute-value components, we formulate a sparse optimization problem as a mixed-integer nonlinear program, whose continuous relaxation is equivalent to the unconstrained minimization of a difference-of-convex function. The approach is applied to Feature Selection in the support vector machine framework, and tested on a set of benchmark instances. Numerical comparisons against both the standard ell _1-based support vector machine and a simple version of the Slope method are presented, that demonstrate the effectiveness of our approach in achieving high sparsity level of the solutions without impairing test-correctness.

Conserved Transcriptional Signatures of Microglia in Health and Disease (P2-3.011)

<h3>Objective:</h3> To determine signatures of microglia in humans and mice in the context of central nervous system (CNS) inflammation and homeostasis. <h3>Background:</h3> Single cell profiling of gene expression has illuminated the heterogeneity of immune cells and accessible phenotypes across organs and under a myriad of conditions. These efforts—typified by recent studies aiming to identify signatures of microglia—typically entail pre-enrichment of cell populations, even from distinct species, using a limited set of surface markers or morphological traits. This step introduces potential bias and subjectivity that may alter the gene expression states that are discerned. <h3>Design/Methods:</h3> We queried the NCBI GEO for datasets of single-cell transcription within CNS tissue from healthy subjects, multiple sclerosis (MS) patients, and mice afflicted with experimental autoimmune encephalomyelitis (EAE). In total, this yielded data on over 135,000 human cells—encompassing a spectrum of lesion chronicity in patients with progressive and relapsing-remitting MS—and 12,000 murine cells. Datasets were pre-processed for quality control and harmonized according to species prior to cross-species integration. We developed classifier models of microglia using logistic regression and support vector machine frameworks. <h3>Results:</h3> Unexpectedly, markers closely linked in the literature to microglia, including Iba-1, were found to be expressed in non-microglial myeloid cells. Other markers previously ascribed to microglia in one species, such as Hexb, failed to represent useful markers in the other species. We identified intra- and inter-species signatures of microglia using machine-learning methods that form the basis of future studies. <h3>Conclusions:</h3> Whole-brain, single-cell gene expression data, unencumbered by bias introduced by prior cell enrichment, enable identification of transcriptional signatures of cell types across species. Signatures of microglia during CNS inflammation identify potentially targetable pathways to understand their role in disease and limit tissue injury. <b>Disclosure:</b> The institution of an immediate family member of Dr. Salinas has received research support from Veterans Health Administration. Navid Manouchehri has nothing to disclose. Rehana Hussain has nothing to disclose. Dr. Stuve has received personal compensation in the range of $0-$499 for serving on a Scientific Advisory or Data Safety Monitoring board for EMD Serono. Dr. Stuve has received personal compensation in the range of $0-$499 for serving on a Scientific Advisory or Data Safety Monitoring board for Novartis. Dr. Stuve has received personal compensation in the range of $0-$499 for serving on a Scientific Advisory or Data Safety Monitoring board for Roche Genentech. Dr. Stuve has received personal compensation in the range of $0-$499 for serving on a Scientific Advisory or Data Safety Monitoring board for TG Therapeutics. Dr. Stuve has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for TG Therapeutics. Dr. Stuve has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for EMD Serono. Dr. Stuve has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for VYNE. Dr. Stuve has received personal compensation in the range of $5,000-$9,999 for serving as an Editor, Associate Editor, or Editorial Advisory Board Member for Therapeutic Advances in Neurological Diseases. Dr. Stuve has received research support from US Department of Veterans Affairs. Dr. Stuve has received research support from National Multiple Sclerosis Society (US). Dr. Stuve has received research support from Exalys. Dr. Stuve has received research support from Merck KGaA.

Machine Learning-Based Unobtrusive Intake Gesture Detection via Wearable Inertial Sensors.

Dietary patterns can be the primary reason for many chronic diseases such as diabetes and obesity. State-of-the-art wearable sensor technologies can play a critical role in assisting patients in managing their eating habits by providing meaningful statistics on critical parameters such as the onset, duration, and frequency of eating. For an accurate yet fast food intake recognition, this work presents a novel Machine Learning (ML) based framework that shows promising results by leveraging optimized support vector machine (SVM) classifiers. The SVM classifiers are trained on three comprehensive datasets: OREBA, FIC, and CLEMSON. The developed framework outperforms existing algorithms by achieving F1-scores of 92%, 94%, 95%, and 85% on OREBA-SHA, OREBA-DIS, FIC, and CLEMSON datasets, respectively. In order to assess the generalization aspects, the proposed SVM framework is also trained on one of the three databases while being tested on the others and achieves acceptable F1-scores in all cases. The proposed algorithm is well suited for real-time applications since inference is made using a few support vector parameters compared to thousands in peer deep neural networks models.

Deep convolutional cross-connected kernel mapping support vector machine based on SelectDropout

Deep neural mapping support vector machine (DNMSVM) has achieved good results in numerous tasks by mapping the input from a low-dimensional space to a high-dimensional space and then using support vector machine for classification. However, it did not consider the connection of different spaces and increased the model parameters. To improve the classification performance while reducing the number of model parameters, we propose a deep Convolutional Cross-connected Kernel Mapping Support Vector Machine framework based on SelectDropout (CCKMSVM-SD). It consists of a feature extraction module and a classification module. The feature extraction module maps the data from low-dimensional to high-dimensional space by fusing the representations of different dimensional spaces through convolutional layers with cross-connections. For some convolutional layers, we use the depthwise separable convolution to replace the original convolution to reduce the number of parameters. Besides, we use SelectDropout to improve its generalization capability. The classification module uses a soft margin support vector machine for classification. The results on three tasks with ten different datasets indicate that CCKMSVM-SD obtains higher classification accuracy than other models with fewer parameters, demonstrating its effectiveness.

PolyFLAG_SVM: a Polymorphic Federated Learning Aggregation of Gradients Support Vector Machines Framework

The critical importance of user privacy in the context of Machine Learning is a hot research topic because it hinders data collection. Consequently, Federated Learning (FL) has emerged as a solution to the privacy problem. Instead of collecting users’ data to train smart models, FL exchanges models with clients, which are trained and sent back to the server for aggregation and global model updating. However, FL still faces some hurdles, such as vulnerability to inference and poisoning attacks. For this reason, this paper proposes PolyFLAG_SVM: Polymorphic Federated Learning Aggregation of Gradients Support Vector Machines Framework. PolyFLAG_SVM is a novel, secure, communication-efficient framework that provides several variants of Support Vector Machines models that follow the Gradient Descent Update technique. The confidence in the security of the proposed model is the result of the polymorphism of the encryption keys used, which guarantees that a cracked or leaked key is useless, since it is not used twice within the FL cycle. Moreover, the proposed framework is communication efficient due to the small size of messages exchanged between servers and clients. The proposed model is explained in detail and evaluated appropriately in this paper.

Risk Assessment for a Twin-Screw Granulation Process Using a Supervised Physics-Constrained Auto-encoder and Support Vector Machine Framework.

Quality risk management is an important task when it pertains to the pharmaceutical industry, as this is directly related to product performance. With the ICH Q9 guidelines, several regulatory bodies have encouraged the pharmaceutical industry to implement risk management plans using scientific and systemic approaches such as quality-by-design to asses product quality. However, the implementation of such methods has been challenging as assessment of risks requires accurate quantitative models to predict changes in quality when variations occur. This study describes a framework that quantitatively assesses risk for a twin screw wet granulation process. This framework consists of a physics-constrained autoencoder system, whose outputs are constrained using physics-based boundary conditions. The latent variables obtained from the auto-encoder are used in a support vector machine-based classifier to understand the granule growth behavior occurring within the system. This framework is able to predict the process outcomes with 86% accuracy and classify the granule growth regimes with a true positive rate of 0.73. Based on the classification the risk associated with the process can be estimated.

A support vector machine framework for fault detection in molecular pump

ABSTRACT The molecular pump fault results in a vacuum leak that can lead to accidents in a fusion device. The fault detection for the molecular pump can provide an early warning to avoid a greater loss in a fusion device. This paper proposes a Support Vector Machine (SVM) framework for fault detection in a molecular pump. This study investigates the key technologies, framework, and system development methods of the data-driven fault detection system for a molecular pump. This study proposes a molecular pump test system to simulate the fault of a molecular pump to solve the highly unbalanced data problem and verify the framework availability. The packaged development framework, implemented in the test system, realizes the fault detection function. The developed system offers better practicability and can effectively detect perform faults in a molecular pump. The results obtained would support new ideas and methods for early warning in molecular pumps and pave the way for stable and reliable vacuum pump systems in a fusion device.

Tool for Predicting College Student Career Decisions: An Enhanced Support Vector Machine Framework

The goal of this research is to offer an effective intelligent model for forecasting college students’ career decisions in order to give a useful reference for career decisions and policy formation by relevant departments. The suggested prediction model is mainly based on a support vector machine (SVM) that has been modified using an enhanced butterfly optimization approach with a communication mechanism and Gaussian bare-bones mechanism (CBBOA). To get a better set of parameters and feature subsets, first, we added a communication mechanism to BOA to improve its global search capability and balance exploration and exploitation trends. Then, Gaussian bare-bones was added to increase the population diversity of BOA and its ability to jump out of the local optimum. The optimal SVM model (CBBOA-SVM) was then developed to predict the career decisions of college students based on the obtained parameters and feature subsets that are already optimized by CBBOA. In order to verify the effectiveness of CBBOA, we compared it with some advanced algorithms on all benchmark functions of CEC2014. Simulation results demonstrated that the performance of CBBOA is indeed more comprehensive. Meanwhile, comparisons between CBBOA-SVM and other machine learning approaches for career decision prediction were carried out, and the findings demonstrate that the provided CBBOA-SVM has better classification and more stable performance. As a result, it is plausible to conclude that the CBBOA-SVM is capable of being an effective tool for predicting college student career decisions.

Classification of macular abnormalities using a lightweight CNN-SVM framework

Macular abnormalities are the main reason for central vision loss, especially in elderly people. Due to global population aging, a heavy burden will be placed on the health care system. Therefore, it is urgent and necessary to develop an automatic and intelligent tool to identify macular abnormalities. Optical coherence tomography is a non-invasive rapid imaging technique to diagnose macular abnormalities. We propose a lightweight convolutional neural network–support vector machine (CNN-SVM) framework consisting of a novel lightweight CNN backbone and an SVM classifier for the accurate detection of macular abnormalities. The CNN-SVM framework achieves excellent performance based on various metrics (precision, recall, F1-score, and accuracy) with an accuracy of 99.8% and demonstrates strong interpretability using heatmap visualization, outperforming several state-of-the-art models (Joint-Attention Network, OpticNet, MobileNet-V3, DenseNet-169, ResNet-50, lesion-aware CNN, Atten-ResNet, least-squares generative adversarial network and others). The proposed CNN-SVM framework is a feasible and reliable tool for the classification of macular abnormalities and shows potential for diagnostic ophthalmology in clinical practice.

Single and Union Non-Parallel Support Vector Machine Frameworks

Single and Union Non-Parallel Support Vector Machine Frameworks

High-dimensional sufficient dimension reduction through principal projections

We develop in this work a new dimension reduction method for high-dimensional settings. The proposed procedure is based on a principal support vector machine framework where principal projections are used in order to overcome the non-invertibility of the covariance matrix. Using a series of equivalences we show that one can accurately recover the central subspace using a projection on a lower dimensional subspace and then applying an ℓ1 penalization strategy to obtain sparse estimators of the sufficient directions. Based next on a desparsified estimator, we provide an inferential procedure for high-dimensional models that allows testing for the importance of variables in determining the sufficient direction. Theoretical properties of the methodology are illustrated and computational advantages are demonstrated with simulated and real data experiments.

Self-training algorithm for hyperspectral imagery classification based on mixed measurement k-nearest neighbor and support vector machine

Self-training is commonly used for hyperspectral imagery semi-supervised learning, which selects confident unlabeled samples and puts them into the labeled set iteratively. However, self-training suffers from the influence of noises and outliers that are wrongly selected as confident samples and guide the accuracy of the classifier going down in iteration. To solve this problem, a self-training algorithm for hyperspectral imagery classification based on a mixed measurement k-nearest neighbor (k-NN) and support vector machine (SVM) framework is proposed. Different from the traditional self-training based on k-NN and SVM, the proposed method utilizes a mixed measurement k-NN to select confident unlabeled samples, and the SVM is applied to help the k-NN method to label the unlabeled samples. The mixed measurement is combined with the spectral distance, spatial distance, and local outlier factor (LOF) distance to explore the feature differences of spectra. Meanwhile, the factors of LOF and k-NN are obtained adaptively, which effectively reduces the computational complexity of the proposed method to select optimized parameters. Experimental results on five hyperspectral images show that the proposed method outperforms the competitors.