论文部分内容阅读
孪生支持向量回归(TSVR)通过快速优化一对较小规模的支持向量机问题获得回归函数.文中提出在原始输入空间中采用Newton法直接优化TSVR的目标函数,从而有效克服TSVR通过对偶二次规划问题求得近似最优解导致性能上的损失.数值模拟实验表明该方法不仅能提高TSVR的性能,并且可降低学习时间.
The Twin Support Vector Regression (TSVR) obtains the regression function by quickly optimizing a pair of smaller SVM problems. In this paper, we propose a Newton method to directly optimize the objective function of TSVR in the original input space so as to effectively overcome the problem of TSVR through dual quadratic programming The approximate optimal solution of the problem leads to the performance loss.The numerical simulation results show that this method can not only improve the performance of TSVR, but also reduce the learning time.