The key problem of training support vector machines is how to solve quadratic programming problem, but for large training examples, the problem is too difficult.

• 训练支持向量机的本质问题就是求解二次规划问题,但对大规模的训练样本来说,求解二次规划问题困难很大。