中文摘要：

针对传统支持向量机无法适应文本数据库不断更新的要求,采用在线最小二乘向量机算法,提出了一种基于LS-SVM的在线文本识别算法,对训练样本进行在线拾取、识别.仿真实验表明该算法计算步骤少、函数收敛速度快,适合文本在线识别.

英文摘要：

The traditional SVM text databases can not meet the requirements of continuously updated with online least-squares algorithm for vector machines,a LS-SVM based on-line text recognition algorithm,to pick up on-line training samples,identification.Simulation steps that the algorithm is less function fast convergence for the text on-line identification.