Based on the equivalence between the original training set and the newly added training set, a new algorithm for SVM-based incremental learning was proposed.

• 基于原训练样本集和新增训练样本集在增量训练中地位等同，提出了一种新的SVM增量学习算法。