The support vector machine (SVM) is a widely used tool in classification problems. SVM trains classifier by solving an optimization problem to decide which instances of the training data set are vectors, necessarily informative form classifier. Since vectors intact tuples taken from set, releasing for public use or shipping clients will disclose private content vectors. This violates privacy-preserving requirements some legal commercial reasons. that learned inherently privacy. privacy...

The effects of varying austenitizing and tempering temperatures on the microstructure, mechanical properties, corrosion resistance 13Cr-2Ni-2Mo martensitic stainless steel were examined. findings indicate that highest hardness optimal are achieved with at 1150°C. This is due to dissolution carbides back into matrix, which results in solid solution strengthening an elevated chromium content within matrix. exhibits when subjected 150°C. As temperature increases, dislocation density matrix...

Multi-Object Multi-Camera Tracking (MOMCT) is essential for real-world applications but remains challenging in low-FPS environments where motion-based tracking ineffective. This study proposes a novel vehicle approach that integrates YOLOv9 object detection, SwinReID feature extraction, and KNN-based matching algorithm to enhance across multiple cameras. Our research leverages the AICUP 2024 Spring Competition dataset, which includes surveillance footage from Chiayi City, along with ImageNet...

Outsourcing the training of support vector machines (SVM) to external service providers benefits data owner who is not familiar with techniques SVM or has limited computing resources. In outsourcing, privacy a critical issue for some legal commercial reasons since there may be sensitive information contained in data. Existing privacy-preserving works are either applicable outsourcing weak security. this paper, we propose scheme without disclosing actual content provider. proposed scheme,...

Fused deposition modeling (FDM) is a form of additive manufacturing where three-dimensional (3D) models are created by depositing melted thermoplastic polymer filaments in layers. Although FDM mature process, defects can occur during printing. Therefore, an image-based quality inspection method for 3D-printed objects varying geometries was developed this study. Transfer learning with pretrained models, which were used as feature extractors, combined ensemble learning, and the resulting model...

We developed a blockchain-based E-marketplace. A decentralized E-marketplace platform utilizing the blockchain technology is implemented. use self-enforcement of smart contracts to secure deposit and process payment. Each transaction verified through recorded ledger. This enables trustless transactions since contract self-executed. The able perform credible without trusted third parties, on are trackable irreversible. Therefore, both buyer seller cannot breach contract. All processes...

Many practical applications of classification require the classifier to produce a very low false-positive rate. Although Support Vector Machine (SVM) has been widely applied these due its superiority in handling high dimensional data, there are relatively little effort other than setting threshold or changing costs slacks ensure In this paper, we propose notion Asymmetric VectorMachine (ASVM) that takes into account false-positives and user tolerance objective. Such new objective formulation...

With the prevalence of mobile devices such as smartphones and tablets, ways people access to Internet have changed enormously. In addition information that can be recorded by traditional Web-based e-commerce like frequent online shopping stores browsing histories, are capable tracking sophisticated behavior. The aim this study is utilize users' behavior reading hotel reviews on subsequently apply text-mining techniques construct user interest profiles make personalized recommendations....

The popularity and development of mobile devices apps have dramatically changed human life. Due to the tremendous still rapidly growing number apps, helping users find that satisfy their demands remains a difficult task. To address this problem, we propose personalized app recommender system based on textual data user reviews store. Topic modeling techniques are applied extract hidden topics reviews, probability distributions utilized represent features apps. Then, profile is constructed...

Consumer reviews play an important role in various e-commerce sites like hotel reservation and app stores. Online consumer are informative because they convey consumers' actual experiences evaluations to the products services received. In this paper, we leverage develop a review-driven recommender service for websites. We semantically explore topics derive product features infer preferences of consumers making recommendations. conduct large-scale experiments on real-world data. The results...

We develop a blockchain-based mobile ticketing system which utilizes smart contracts and multi-signature to enforce authorize the executions of transactions. The transactions buying using tickets are verified by blockchain, immutable ledger on blockchain ensures authenticity security tickets. experimental results show that proposed is very efficient requires only nominal cost.

Support vector machine (SVM) is a widely used tool in classification problem. SVM solves quadratic optimization problem to decide which instances of training dataset are support vectors, i.e., the necessarily informative form classifier. The vectors intact tuples taken from dataset. Releasing classifier public use or shipping clients will disclose private content violating privacy-preservation requirement some legal commercial reasons. To best our knowledge, there has not been work extending...

The breast ultrasound is not only one of major devices for tissue imaging, but also important methods in tumor screening. It non-radiative, non-invasive, harmless, simple, and low cost American College Radiology (ACR) proposed the Breast Imaging Reporting Data System (BI-RADS) to evaluate far more lesion severities compared traditional diagnoses according five-criterion categories masses composition described as follows: shape, orientation, margin, echo pattern, posterior features. However,...

Previous chapter Next Full AccessProceedings Proceedings of the 2011 SIAM International Conference on Data Mining (SDM)Efficient Kernel Approximation for Large-Scale Support Vector Machine ClassificationKeng-Pei Lin and Ming-Syan ChenKeng-Pei Chenpp.211 - 222Chapter DOI:https://doi.org/10.1137/1.9781611972818.19PDFBibTexSections ToolsAdd to favoritesExport CitationTrack CitationsEmail SectionsAboutAbstract Training support vector machines (SVMs) with nonlinear kernel functions large-scale...

The Support Vector Machines (SVMs) have been widely used for classification due to its ability give low generalization error. In many practical applications of classification, however, the wrong prediction a certain class is much severer than that other classes, making original SVM unsatisfactory. this paper, we propose notion Asymmetric Machine (ASVM), an asymmetric extension SVM, these applications. Different from existing extensions such as thresholding and parameter tuning, ASVM employs...

The systems modeling technique is used for developing an abstract model to help analyze and understand the functionality of a system. In this paper, we introduce new approach general-purpose called structure-behavior coalescence (SBC), which supports integration both structural behavioral aspects system in single diagram. Process algebra utilized describe interactions between components SBC combines graphical notation process provide formalism intuitive representation development conceptual...

Kernel k-means is a useful way to identify clusters for nonlinearly separable data. Solving the kernel problem time consuming due quadratic computational complexity. Outsourcing computations of solving external cloud computing service providers benefits data owner who has only limited resources. However, privacy critical concern in outsourcing since may contain sensitive information. In this paper, we propose method privacy-preserving based on randomized matrix. The experimental results show...

The financial media news reports have helped market participants to know about the strength and weakness of corporate future development, they still encounter a new challenge when analyzing businesses, because their velocity, volume, variety, veracity (4V). In comparison with numerical ratios derived from reports, we believe textual information provides much more essential valuable cues pertaining existence difficulties, preparations satisfy some selective accounting principle or estimation...

We explore in this paper a novel sampling algorithm, referred to as algorithm PAS (standing for proportion approximation sampling), generate high-quality online sample with the desired rate. The quality refers consistency between population and of each categorical value database. Note that state-of-the-art preserve has examine pilot priori is thus not applicable incremental mining applications. To remedy this, adaptively determines inclusion probability incoming tuple such way can be...

Due to advances in medicine, the societies have become aging, which leads problem of missing teeth. Missing teeth negatively affects not only external appearance but also personal dietary habits. Therefore, it has an important issue dentistry solve In Taiwan, dental prostheses for are produced by technicians laboratories. The traditional production is labor consuming. But need human power laboratories will be satisfied future due low birth rate. necessary decrease this area strategies...

Unified Modeling Language (UML) possesses the model multiplicity problems because it does not integrate systems structure with behavior. To overcome this shortcoming, we propose an approach to enrich UML from singularity by creating its kernel structure-behavior coalescence process algebra. The and behavior models will be derived model. characteristics of single prevents problem.

Due to recent explosion of text data, researchers have been overwhelmed by ever-increasing volume articles produced different research communities. Various scholarly search websites, citation recommendation engines, and databases created simplify the tasks. However, it is still difficult for be able identify potential topics without doing intensive reviews on a tremendous number published journals, conferences, meetings, workshops. In this paper, we consider novel topic diffusion discovery...