基于增量学习的直推式支持向量机算法

针对直推式支持向量机在进行大数据量分类时出现精度低、学习速度慢和回溯式学习多的问题，提出了一种基于增量学习的直推式支持向量机分类算法，将增量学习引入直推式支持向量机，使其在训练过程中仅保留有用样本而抛弃无用样本，从而减少学习时间，提高分类速度。实验结果表明，该算法具有较快的分类速度和较高的分类精度。

Transductive support vector machines based on incremental learning

Aiming at the problem of lower precision, slower training speed and more back learning steps when transducitve support vector machine learning algorithm carry on a great deal of data classification, a new transductive support vector machine based on incremental learning was proposed. The incremental learning was introduced into transductive support vector machine; thereby, the algorithm only employed the useful samples and discarded unwanted samples in the training process, reduced the study time and improved the classification speed of the algorithm. The experimental results show that the algorithm is of faster classification speed and higher classification accuracy.