摘要
小样本分类问题由于包含的训练样本个数比较少,通常不足以训练一个理想的分类模型。小样本分类问题普遍存在于现实世界中,因此提高分类器在小样本分类问题上的性能就成为了研究的热点。本文针对该问题,提出了一种添加均匀分布噪声的数据扰动小样本分类算法。该算法首先对每一个原始样本添加一个服从均匀分布的噪声,对原始数据进行一定程度的扰动。然后在所获得的扰动数据集上训练分类模型。在UCI标准数据集上的仿真实验表明,本文算法较传统分类方法,能更有效地提高小样本分类问题的分类性能。
Due to the relatively small number of training samples contained in small sample classification problems,it is insufficient to train a good classification model.As the small sample classification problems occur frequently in the real world,so how to improve the performance of classifiers on small sample classification problems becomes a hot-point.For this problem,in this paper,we propose a data disturbance small sample classification algorithm by adding uniform distribution noise.Firstly,we add a noise following uniform distribution to the original samples,causing a disturbance to some extent.Then train a classification model in the disturbed data set.The simulation experiments on UCI standard data sets show that the proposed algorithm can effectively improve the classification performance on small sample classification problems compared with traditional classification methods.
出处
《科技通报》
北大核心
2013年第6期122-124,共3页
Bulletin of Science and Technology
关键词
小样本分类
均匀分布
噪声
数据扰动
small sample classification
uniform distribution
noise
data disturbance
