首页|Investigators from University of Bejaia Have Reported New Data on Support Vector Machines (An Efficient Primal Simplex Method for Solving Large-scale Support Ve ctor Machines)
Investigators from University of Bejaia Have Reported New Data on Support Vector Machines (An Efficient Primal Simplex Method for Solving Large-scale Support Ve ctor Machines)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News – Current study results on Support Vecto r Machines have been published. According to news reporting out of Bejaia, Alger ia, by NewsRx editors, the research stated, “In the last two decades, active set methods for training support vector machines (SVMs) have received a lot of atte ntion due to their robustness and have shown excellent results for training larg e-scale problems. In this paper, we propose the primal simplex method to solve t he quadratic programming problem encountered during the training phase of an SVM classification problem.” Our news journalists obtained a quote from the research from the University of B ejaia, “Our novel approach, named PSM-SVM, generates iteratively a decreasing se quence of feasible points that converges to the optimal solution. Contrary to ex isting active set algorithms, PSM-SVM has the particularity to avoid using the n ull -space matrix and also the reduced Hessian matrix when the descent direction is calculated at each iteration. In addition, it starts with a workingset havin g only one support vector and also guarantees the nonsingularity of the basic ma trix during all the iterations process. We have theoretically proven its global convergence and calculated its computational complexity.”
BejaiaAlgeriaEmerging TechnologiesMachine LearningSupport Vector MachinesVector MachinesUniversity of Bejaia
NETL
NSTL
