为提高基于超球的支持向量机算法中样本数据较多时的训练速度,提出一种构造最小超球的并行融合算法.该算法将全部训练数据集依据特定策略分割成若干个子数据集,分别对各个子数据集进行训练,对所得到的各子数据集的支持向量与融合数据进行训练,构造最小超球.仿真结果表明,并行融合算法在保证分类精度的情况下,能够显著减少训练时间,提高效率,且支持向量的数目较少.同时也验证了该文对Gauss核函数分析的正确性. In order to improve the training speed of hypersphere-based SVM algorithm, a parallel fusion algorithm for constructing the minimum hypersphere is proposed, which divides all training data sets into several sub-datasets according to specific strategies, Training of each sub-dataset and training the resulting support vector and fusion data of each sub-dataset to construct the minimum hypersphere.The simulation results show that the parallel fusion algorithm can significantly reduce the training time under the condition of ensuring the classification accuracy, Improve the efficiency, and the number of support vectors is less. At the same time, the correctness of Gauss kernel function analysis is verified.