Lagrange支持向量回归机算法研究

Research on lagrange support vector regression arithmetic

支持向量回归机问题的研究远没有像支持向量机问题成熟完善,支持向量回归机对函数拟合(回归逼近)具有重要的理论和应用意义。借鉴分类问题的有效算法,将其推广到回归问题中来,针对Lagrange支持向量机(LSVM)算法,提出了有效的Lagrange支持向量回归机(LSVR)算法,在若干不同维数的数据集上,对LSVR算法、ASVR算法和LibSVM算法进行数值试验,并进行比较分析。数值试验表明LSVR算法是有效的,与当前流行的求解支持向量回归机的算法相比,在时间和正确度上都有一定的优势。

The research on support vector regression is not mature and perfect like the support vector machine. Support vector regression has an important theoretical and applicable significance on function regression （regression approximation）. Using effective arithmetic of classifier for reference, it is extended to the matter ofregression. Aimed at Lagrange support vector machine arithmetic, The effective Lagrange support vector regression arithmetic is put forward. On several different data aggregation of dimensions, the numerical value experiment and comparison are carried out on LSVR arithmetic, ASVR arithmetic and LibSVM arithmetic. The numerical value test has improved that the LSVR arithmetic is effective and it has LSVR arithmetic solution.