S. Asharaf, M. Narasimha Murty, Shirish Shevade · 2007
Even though several techniques have been proposed in the literature for achieving multiclass classification using Support Vector Machine(SVM), the scalability aspect of these approaches to handle large data sets still needs much of exploration. Core Vector Machine(CVM) is a technique for scaling up a two class SVM to handle large data sets. In this paper we propose a Multiclass Core Vector Machine(MCVM). Here we formulate the multiclass SVM problem as a Quadratic Programming(QP) problem defining an SVM with vector valued output. This QP problem is then solved using the CVM technique to achieve scalability to handle large data sets. Experiments done with several large synthetic and real world data sets show that the proposed MCVM technique gives good generalization performance as that of SVM at a much lesser computational expense. Further, it is observed that MCVM scales well with the size of the data set.
Cited by 42
S. Asharaf, M. Narasimha Murty, Shirish Shevade · Proceedings · 2006
Core vector machine(CVM) is suitable for efficient large-scale pattern classification. In this paper, a method for improving the performance of CVM with Gaussian kernel function irrespective of the orderings of patterns belonging to different classes within the data set is proposed. This method employs a selective sampling based training of CVM using a novel kernel based scalable hierarchical clustering algorithm. Empirical studies made on synthetic and real world data sets show that the proposed strategy performs well on large data sets.
Cited by 11
Asharaf, S, Murty, Narasimha M, Shevade, SK · NOT FOUND REPOSITORY (Indian Institute of Science Bangalore) · 2006
Core Vector Machine(CVM) is suitable for efficient large-scale pattern classification. In this paper, a method for improving the performance of CVM with Gaussian kernel function irrespective of the orderings of patterns belonging to different classes within the data set is proposed. This method employs a selective sampling based training of CVM using a novel kernel based scalable hierarchical clustering algorithm. Empirical studies made on synthetic and real world data sets show that the proposed strategy performs well on large data sets.
Cited by 4
Yongqing Wang, Niu Xiao-tai, Liang Chang · 2010
Traditional methods for solving multi-class problems, well-known as multi-SVMs, always combine certain decomposed binary-SVMs' results to formulate the final decision function. The prevalent methods are `one vs. one' and `one vs. all', which are based on a voting scheme among the binary classifiers to derive the winning class. However, they do not scale well with the data size and class number. Core Vector Machine (CVM) is a promising technique for scaling up a binary-SVM to handle large data sets with the greedy-expansion strategy, where the kernels are required to be normalized to ensure the equivalence between the kernel-induced spaces of SVM and Minimum Enclosing Ball (MEB). The idea proposed by CVM can also be utilized to formulate multi-SVM to MEB, by which we propose an approximate MEB algorithm with smaller core sets to handle multi-SVM. The experimental results on synthetic and benchmark data sets demonstrate the competitive performances of the method we proposed both on training time and training accuracy.
Cited by 4
Moritz Heusinger, Frank-Michael Schleif · Lecture notes in computer science · 2023
Cited by 0
Rahul Kumar, Saumil Maheshwari, Apoorva Mishra et al. · 2017
We are in an era where scaling up the processing speed of computers by increasing the clock speed of processors had become an ineffective strategy for handling Big Data. Now, computers have increased number of processors embedded within, which has motivated various programmer to take advantage of these architectures. To increase the processing speed, parallelization of algorithms can be used as an alternative rather than optimizing the existing algorithms. In the paper, we have deployed a parallel programming framework that could take advantage of multi-core structure of processors. This framework can be used for different learning algorithms that satisfies Statistical Query Model (SQM) i.e., can be simplified in “Summation form”. We have used Google’s MapReduce model for achieving parallelization in learning algorithms on two datasets (Ionosphere, Breast Cancer datasets) from UCI machine learning repository. Our implemented model gives better result than existing implementations and also verifies the linear speedup of processing with an increase in number of cores in the processors.
Cited by 1
Di Wang, Bo Zhang, Peng Zhang et al. · Pattern Recognition · 2010
Cited by 22
Jianhua Xu · Pattern Recognition · 2014
Cited by 22
Babak Afshin, Mohammad Ebrahim Shiri, Kamran Layeghi et al. · Neural Processing Letters · 2023
Cited by 1
Chengquan Huang, Fu-Lai Chung, Shitong Wang · Neural Networks · 2015
Cited by 32
Junchen Si, Yuanyuan Wang, Yongchang Guan et al. · Journal of Physics Conference Series · 2024
Abstract The development of smart grids requires enhanced data integration, robust risk assessment, and dynamic response optimization. In this paper, a multi-core learning Support Vector Machine (SVM) model is presented to improve the accuracy and efficiency of load and photovoltaic output forecasting. The model leverages kernel function optimization and parallel computing frameworks to handle large-scale data efficiently. Additionally, a comprehensive risk assessment system is developed to quantify risks such as overvoltage, undervoltage, line overload, and load loss in distribution networks. An adaptive genetic algorithm-based risk control model is also proposed, optimized in two stages—day-ahead and intra-day—to achieve minimal comprehensive risk through real-time adjustments in distributed power output and electric vehicle charging strategies. Furthermore, an integrated virtual synchronous control online verification method for source-network-load-storage is introduced, enhancing system response speed and control accuracy. These innovations collectively provide a solid theoretical foundation and technical support for the efficient and safe operation of smart grids, addressing the increasing demands of modern energy systems.
Cited by 1Open access
Xusheng Gan, Duanmu Jingshun, Jiafu Wang et al. · Knowledge-Based Systems · 2012
Cited by 92
Yueyao Li, Chenglong Bao, Wenxun Xing · ArXiv.org · 2025
Support vector machine (SVM) is a well known binary linear classification model in supervised learning. This paper proposes a globalized distributionally robust chance-constrained (GDRC) SVM model based on core sets to address uncertainties in the dataset and provide a robust classifier. The globalization means that we focus on the uncertainty in the sample population rather than the small perturbations around each sample point. The uncertainty is mainly specified by the confidence region of the first- and second-order moments. The core sets are constructed to capture some small regions near the potential classification hyperplane, which helps improve the classification quality via the expected distance constraint of the random vector to core sets. We obtain the equivalent semi-definite programming reformulation of the GDRC SVM model under some appropriate assumptions. To deal with the large-scale problem, an approximation approach based on principal component analysis is applied to the GDRC SVM. The numerical experiments are presented to illustrate the effectiveness and advantage of our model.
Cited by 0Open access
Wenjun Hu, Fu-Lai Chung, Shitong Wang · Neural Networks · 2011
Cited by 11
Saurabh Kumar Gupta, Satish B. Satpal, Sauvik Banerjee et al. · International Journal on Smart Sensing and Intelligent Systems · 2016
Abstract Honeycomb sandwich structures are extensively used in aerospace, aeronautic, marine and automotive industries due to their high strength-to-weight ratios, high energy absorption capability and effective acoustic insulation. Unfortunately, either presence of disbond along the skin-core interface or emergence of disbond due to repeated loading, aging or an intensive load can jeopardize the integrity and safety of the whole structure The current work presents a new array based technique for health monitoring of these structures using support vector machine (SVM). The proposed technique is first used on simulated mode shape data of the structure and then the technique is validated using experimental mode shape data. The experimental set up has been developed in laboratory and Laser Doppler Vibrometer (LDV) is used to extract the experimental mode shapes. The results have been obtained using both support vector classification and regression analysis and it is found that that the former is better at prediction of debond location
Cited by 2Open access
Radhika V. Menon, Indrajit Chakrabarti · Microprocessors and Microsystems · 2023
Cited by 3
Lingye Kong, Mengyuan An · 2024
Aiming at the problem of low efficiency and insufficient accuracy of existing anomaly detection methods, this paper proposes a new anomaly detection model based on autoencoder and second-order optimized core vector machine (CVM_2o). First, the VAE model is used to learn the distribution characteristics and potential structure of the input data to generate an ideal data distribution. Secondly, this paper proposes a second-order bidirectional conditional gradient method, which is integrated into the traditional CVM algorithm to construct CVM_2o model, so as to achieve accurate and fast anomaly data detection. Finally, the numerical experimental results show that compared with the traditional anomaly detection method, this method has less computational time under the premise of ensuring the detection accuracy, which proves the effectiveness and practicability of the VAE-CVM_2o model in the field of anomaly detection.
Cited by 0
Želimir Galić, Matija Varga, Neven Bulić et al. · 2024
This paper presents the implementation of vector control and highlights the capabilities of the AM2634 microcontroller, specifically designed for executing time-critical algorithms for the control of modern electrical drives. The development process involved the integration of the TMDSCNCD263 development kit, tailored for rapid prototyping and peripheral connectivity with this new generation of microcontrollers. A custom printed circuit board (PCB) serving as an interface for TMDSCNCD263 was developed to achieve interaction with the rest of the industrial drive system. The developed control system enables comprehensive testing and validation of various control algorithms for induction machine fed with a low-voltage industrial power converter. The custom PCB was designed to meet specific industrial requirements and to ensure fiber optic connections with IGBT drivers. Subsequent laboratory tests were conducted to validate the effectiveness of the algorithms and to verify the functionality of the developed system. Finally, the results obtained are presented.
Cited by 0
Nan Li, Zhonghua Sun, Fan Jiang · BMC Bioinformatics · 2008
By improving both the descriptions of the interface residues and their surrounding environment and the training strategy, better SVM models were obtained and shown to outperform previous methods. Our tests on the unbound protein structures suggest further improvement is possible.
Cited by 74Open access
Xiu Li, Jin Yu, Zhuo Jia et al. · 2015
Support vector machine (SVM) and its derivative algorithms have been increasingly used to predict algal blooms recently. However, its computation complexity remains an annoying problem. To improve the time cost of SVM, a hybrid approach is proposed in this paper based on Partial Least Square (PLS) feature extraction and Core Vector Machine Regression (CVR) algorithm. We describe the principle of our algorithm and the implementation steps in detail. Based on the biweekly data gathered from Tolo Harbour, Hong Kong, some comparative analysis of the performance of PLS-CVR and other algorithms are presented. We develop these prediction models with different lead time (7-day and 14-day) to study further. The results indicate that the use of biweekly data can simulate the general trend of algal biomass reasonably. The experimental results show that our algorithm can reduce the time cost significantly compared to conventional SVM algorithm while maintaining a satisfactory accuracy.
Cited by 1