摘要
基于经典统计学的机器学习算法,在解决小样本学习问题时表现得不能令人满意。在总结分析小样本机器学习算法特点的基础上,以支持向量机(SVM)学习算法为例,定量分析了影响其泛化性能、学习性能的几个因素,实验结果与理论分析结论取得了良好的一致性;SVM用于解决KTH-TIPS纹理图像分类问题,取得了很好的实验结果。
Some limitation of the machine learning algorithm based on the classical statistics has been displayed when it is used to solve the learning problems with limited samples. On the foundation of summarizing their characteristics, the quantitative analysis for generalization function and learning function are presented in this paper, taking example for the support vector machine （SVM） algorithm. The consistency between experimental result and theoretical conclusion is perfect, and a favorable classification result has been gained when SVM is used to KTH-TIP texture images.
出处
《海洋测绘》
2007年第3期16-19,共4页
Hydrographic Surveying and Charting
关键词
图像处理
机器学习
统计学习理论
支持向量机
纹理图像
image processing
machine learning
statistical learning theory
support vector machine
texture image
