针对传统SVC方法在样本容量大时存在训练时间过长、精度不高等不足,建立了一种变量可分离的支持向量分类模型DCSVC及算法,并应用于随机函数生成数据分类学习及戈尾属植物数据集分类预测中,从理论与实践上证明了DCSVC算法优于传统SVC算法(分类正确率较高而且训练时间较短).
To solve the disadvantage of long training time and low accuracy of traditional SVC model with big study sample data, sets up a new kind of support vector classification machine model "DCSVC" whose variables can be separated. The IRIS standard test data set and random data set experiment demonstrates that the accuracy of DCSVC is higher than that of traditional support vector machine and the training time is lower.