论文部分内容阅读
针对支持向量机在对海量训练样本进行训练时,训练速度慢而导致难以应用的问题,通过分析训练样本数目与训练时间之间的关系,利用支持向量机对小样本学习的良好特性,提出了基于样本分组的支持向量机快速训练算法。将海量样本分成小样本进行训练,然后对训练得到的多个支持向量机进行加权处理得到决策函数。此方法在标准数据以及陀螺仪参数漂移数据上进行了仿真应用,方针结果证明该方法可大幅提高训练速度,同时保证了较好的泛化能力。
In order to solve the problem that SVM is difficult to apply when training massive training samples, the training speed is slow and difficult to apply. By analyzing the relationship between the number of training samples and training time and using the good features of SVM for small sample learning, Support Vector Machine Fast Training Algorithm Based on Sample Grouping. The massive samples are divided into small samples for training, and then the training of multiple support vector machines weighted processing decision-making function. The method is applied to the simulation of standard data and gyro parameter drift data. The results show that this method can greatly improve the training speed and ensure the good generalization ability.