基于向量投影的支撑向量预选取

支撑向量机是近年来新兴的模式识别方法，在解决小样本、非线性及高维模式识别问题中表现出了突出的优点．但在支撑向量机中，支撑向量的选取相当困难，这也成为限制其应用的瓶颈问题．该文对支撑向量机的机理经过认真分析，研究其支撑向量的分布特性，在不影响分类性能的前提下，提出了基于向量投影的支撑向量预选取法，从训练样本中预先选择具有一定特征的边界向量来代替训练样本进行训练，这样就减少了训练样本，大大加快了支撑向量机的训练速度。

Pre -extracting Support Vector for Support Vector Machine Based on Vector Projection

Support Vector Machine (SVM), a novel method of the pattern recognition, presents excellent performance in solving the problems with small sample, nonlinear and local minima. However, training a support vector machine (SVM) is equivalent to solving a linearly constrained quadratic programming (QP) problem in a number of variables equal to the number of data points. This optimization problem is known to be challenging when the number of data points exceeds few thousands. Also, it is well known that the ratio of support vectors (SVs) is far low in many practical circumstances. So the method of pre extracting SVs to train classifier becomes a novel task in SVM field. In this paper, on a deep investigation into the principle of SVM and its characteristic, we a new method for pre extracting SVs based on vector projection is introduced, which reduces the training samples greatly and speeds up the SVM learning, while the ability of SVM remains unchanged.