×

Lavrentiev regularization method in rank learning. (Chinese. English summary) Zbl 1424.68125

Summary: This paper studies the convergence of Lavrentiev regularization method in ranking learning. To solve the pairwise ranking problem, the main existing method is to solve a minimization problem with Tikhonov regularization. However, applying the self conjugation operator with the standard Tikhonov regularization scheme can be transformed into a form of that with Lavrentiev regularization scheme. Moreover, the convergence order of this Lavrentiev sorting algorithm under a certain smoothness condition is obtained through a convergence analysis of the method.

MSC:

68T05 Learning and adaptive systems in artificial intelligence
65F22 Ill-posedness and regularization problems in numerical linear algebra
PDFBibTeX XMLCite