近似支持向量机参数选择及稀疏度的研究-应用数学专业论文.docxVIP

I I 摘 要 摘 要 标准支持向量机的求解属于二次规划问题，对于大样本集的问题, 支持向量机的训 练时间较长。为了降低时间复杂度，前人提出了近似支持向量机。其中标准的近似支持 向量机是一种新型的求解分类问题的方法，它不是二次规划问题，而是一个正则最小二 乘问题能求出解析解，因此训练速度大大提高。 本文针对数据不平衡问题，提出了一种基于密度加权的近似支持向量机方法，为每 一个样本分配不同的隶属度，首先计算每一个样本的密度信息，根据样本的密度信息对 样本的误差项进行密度加权。由于每一个样本的密度指标值不同，所以样本对模型的贡 献也是不同的，密度加权近似支持向量机给不同的样本赋予了不同的惩罚因子。近似支 持向量机使所有的样例都成为支持向量，使得稀疏度无法控制，因此本文给出了一种控 制稀疏度的方法，有效地控制了密度加权近似支持向量机的稀疏度。核参数的优劣直接 影响分类器泛化能力的好坏，给出了一种核参数选择的方法，使得分类器具有更好的分 类性能。将该方法用于 UCI 数据集中，分别与其他方法进行比较，实验结果表明增量密 度加权近似支持向量机具有较好的分类性能。 关键词 支持向量机 近似支持向量机 密度 增量 稀疏度 II II Abstract Abstract The solving of standard support vector machine is a quadratic programming problem. It needs long training time for large sample sets. In order to reduce the time complexity, many scholars have put forth on proximal support vector machine. The standard proximal support vector machine is a new type of solving classification problem, it is not a quadratic programming problem, but a regular least squares problem, and it can obtain the analytical solution, so it improves the training speed. This paper presents a proximal support vector machine based on the density-weighted method. It assigns a different degree for each sample, and it gives the density weight for the sample error according to density information of the sample. Proximal support vector machine make all sample become support vector, so the sparsity can not control. This paper gives a method of controlling the sparsity, and it effectively controls the sparsity. The quality of the kernel parameter directly affect the generalization capability of the classifier, this paper presents a kernel parameter selection method, so the classifier has better classification performance.This method is used in UCI data set, and it is compared with other methods respectively. The experimental results show that the incremental density weighted proximal support vector machines have better classification performance. Keywords Support Vector Machine Proximal Support Vector Machine Density Increment Sparse