Large Scale Ranking Using Stochastic Gradient Descent

Authors

  • Engin Tas Department of Statistics, Faculty of Science and Literature, Afyon Kocatepe University, Turkey

DOI:

https://doi.org/10.7546/CRABS.2022.10.03

Keywords:

ranking, stochastic gradient descent, east-squares, gradient methods, stochastic optimization

Abstract

A system of linear equations can represent any ranking problem that minimizes a pairwise ranking loss. We utilize a fast version of gradient descent algorithm with a near-optimal learning rate and momentum factor to solve this linear equations system iteratively. Tikhonov regularization is also integrated into this framework to avoid overfitting problems where we have very large and high dimensional but sparse data. 

Author Biography

Engin Tas, Department of Statistics, Faculty of Science and Literature, Afyon Kocatepe University, Turkey

Mailing Address:
Department of Statistics,
Faculty of Science and Literature,
Afyon Kocatepe University
Afyonkarahisar, 03200, Turkey

E-mail: engintas@aku.edu.tr

Downloads

Published

30-10-2022

How to Cite

[1]
E. Tas, “Large Scale Ranking Using Stochastic Gradient Descent”, C. R. Acad. Bulg. Sci., vol. 75, no. 10, pp. 1419–1427, Oct. 2022.

Issue

Vol. 75 No. 10 (2022)

Section

Mathematics

License

Copyright (c) 2022 Proceedings of the Bulgarian Academy of Sciences

Copyright (c) 2022 Proceedings of the Bulgarian Academy of Sciences

Copyright is subject to the protection of the Bulgarian Copyright and Associated Rights Act. The copyright holder of all articles on this site is Proceedings of the Bulgarian Academy of Sciences. If you want to reuse any part of the content, please, contact us.