论文部分内容阅读
鉴于传统支持向量机分类过程的计算量和支持向量的个数成正比,为了提高分类决策的速度,提出一种约简支持向量的快速分类算法,该算法对原始的支持向量进行特定比例的模糊均值聚类操作,按照分类误差最小的原则构建最小线性二乘回归模型,求解新的支持向量系数和决策函数的偏置.人造数据集和标准数据集上的实验表明,约简50%支持向量后,可以在保持分类精度在无统计意义的明显损失的前提下,使得分类速度提高50%.
In view of the fact that the traditional SVM classification process is directly proportional to the number of support vectors, in order to improve the speed of classification decision, a fast classification algorithm of reduction support vector is proposed, which performs a certain proportion of fuzzy obfuscation on the original support vector Mean clustering operation, the minimum linear regression model is constructed according to the principle of minimum classification error to solve the new support vector coefficient and bias of decision function.Experiments on artificial datasets and standard datasets show that the reduction of 50% support vector After the classification accuracy can be maintained in the absence of significant loss of statistical significance, making the classification speed increased by 50%.