Incremental learning is one of the most effective methods accumulated data and large-scale data. The newly increased samples previously known works on incremental are usually independent identically distributed. To study how dependent sampling influence ability support vector machines (ISVM) algorithm, in this paper we introduce an ISVM based Markov resampling (MR-ISVM), give experimental research MR-ISVM algorithm. results indicate that algorithm has not only smaller misclassification rates...

With the increasing prevalence of mobile devices, people prefer to use smartphones make payments, take photos, and collect personal vital information. Due high possibility smartphone illegal access, security privacy devices become more important critical. In this article, we present FusionAuth, a sensor-based continuous authentication system leveraging accelerometer, gyroscope, magnetometer on capture users' behavioral patterns. order improve performance enhance reliability, are among first...

The previously known works studying the generalization ability of support vector machine classification (SVMC) algorithm are usually based on assumption independent and identically distributed samples. In this paper, we go far beyond classical framework by SVMC uniformly ergodic Markov chain (u.e.M.c.) We analyze excess misclassification error u.e.M.c. samples, obtain optimal learning rate for also introduce a new sampling to generate samples from given dataset, present numerical studies...

In this paper, we consider online support vector machine (SVM) classification learning algorithms with uniformly ergodic Markov chain (u.e.M.c.) samples. We establish the bound on misclassification error of an SVM algorithm u.e.M.c. samples based reproducing kernel Hilbert spaces and obtain a satisfactory convergence rate. also introduce novel sampling, present numerical studies ability sampling for benchmark repository. The show that performance is better than classical random as size...

Heterogeneous change detection (CD) is widely applied in various fields such as urban planning, environmental monitoring, and disaster management. It enhances the accuracy comprehensiveness of surface monitoring by integrating multi-sensor remote sensing data. Scholars have proposed many graph-based methods to address issue incomparable heterogeneous images caused imaging differences. However, these often overlook influence changes vertex status on graph structure, which limits their ability...

Unsupervised graph-structure learning (GSL) which aims to learn an effective graph structure applied arbitrary downstream tasks by data itself without any labels' guidance, has recently received increasing attention in various real applications. Although several existing unsupervised GSL achieved superior performance different analytical tasks, how utilize the popular masked autoencoder sufficiently acquire supervision information from for improving effectiveness of learned been not...

Fisher linear discriminant (FLD) is a well-known method for dimensionality reduction and classification that projects high-dimensional data onto low-dimensional space where the achieves maximum class separability. The previous works describing generalization ability of FLD have usually been based on assumption independent identically distributed (i.i.d.) samples. In this paper, we go far beyond classical framework by studying Markov sampling. We first establish bounds performance uniformly...

To solve the problems of susceptibility to image noise, subjectivity training sample selection, and inefficiency state-of-the-art change detection methods with heterogeneous images, this study proposes a post-classification method for images improved hierarchical extreme learning machine (HELM). After smoothing suppress selection is defined train HELM each image, in which feature extraction respectively implemented parameters need not be fine-tuned. Then, multi-temporal maps extracted from...

The objective of this paper is to develop a new effective method for hierarchical segmentation multitemporal ultrafine-beam synthetic aperture radar (SAR) data in urban areas. Multitemporal RADARSAT-2 highresolution horizontal transmit and receive-Synthetic Aperture Radar (HH-SAR) images acquired the rural-urban fringe Greater Toronto Area during summer 2008 are selected research. Stationary wavelet transform (SWT) algebraic multigrid (AMG) proposed SAR data. SWT applied decomposition image...

Ultrasensitive room temperature real-time NO₂ sensors are highly desirable due to potential threats on environmental security and personal respiratory. Traditional gas with operated temperatures (200-600 °C) limited reversibility mainly constructed from semiconducting oxide-deposited ceramic tubes or inter-finger probes. Herein, we report the functionalized graphene network film assembled an electrospun three-dimensional (3D) nanonetwork skeleton for ultrasensitive sensing. The functional 3D...

Support vector machine (SVM) is one of the most widely used learning algorithms for classification problems. Although SVM has good performance in practical applications, it high algorithmic complexity as size training samples large. In this paper, we introduce (SVMC) algorithm based on k-times Markov sampling and present numerical studies SVMC with benchmark data sets. The experimental results show that not only have smaller misclassification rates, less time training, but also obtained...

This paper considers the generalization ability of two regularized regression algorithms [least square (LSRR) and support vector machine (SVMR)] based on non-independent identically distributed (non-i.i.d.) samples. Different from previously known works for non-i.i.d. samples, in this paper, we research bounds uniformly ergodic Markov chain (u.e.M.c.) Inspired by idea Monto Carlo (MCMC) methods, also introduce a new sampling algorithm to generate u.e.M.c. samples given dataset, then, present...

Functional linear regression is one of the main modeling tools for working with functional data. Since data are usually stream essentially and there some noises in Many numerical research studies machine learning indicate that noise samples not only increase amount storage space, but also affect performance algorithm. Therefore, this paper we consider a new strategy by introducing incremental learning, Markov sampling propose novel square algorithm based on (FILSR-MS). To have better...