不平衡样本集中SVM的应用综述

A SURVEY OF SVM IN UNBALANCED DATA SETS

支持向量机是由Vapnik等人于上世纪90年代提出的一种崭新的学习机器,它作为统计学习理论的实现方法,是处理小样本学习的有效工具。支持向量机克服了神经网络收敛速度慢、解不稳定、泛化性差的缺点,在模式识别、信号处理、自动化、通讯等领域得到了广泛应用。在不平衡样本集中,样本数量上的差异导致不同类别的样本对于训练算法提供的信息不对称,所以很多性能较好的算法用来处理不平衡的样本集时得不到令人满意的效果。很多的科研人员对该问题进行了广泛而深入的研究,较为系统地回顾了这一个研究分支在过去10年的发展动态。

Support Vector Machine（SVM） was proposed by Vapnik et. al. in 1990＇s. SVM is a new and outstanding learning machine and is an efficient machine-learning tool in dealing with small samples. SVM overcomes some shortcomings of neural network, such as slow convergence, unstable solution, and bad generalization ability. So it has been widely applied to many areas, such as pattern recognition, signal processing, automation, communication, etc. In the unbalanced sets, the difference of the sample quantities of different classes leads to the asymmetry of the information provided by the different classes used in the training algorithm. Many excellent algorithms haven＇t shown the distinguished performances when applied to the unbalanced sets. Many scholars have done deep researches on this problem to improve the performance of SVM. The advances in such algorithm studies in the last ten years are reviewed.