Matrix completion under interval uncertainty
详细信息 查看全文
- 作者：Jakub Mareček ; a ; jakub.marecek@ie.ibm.com" class="auth_mail" title="E-mail the corresponding author ; Peter Richtá ; rikb ; peter.richtarik@ed.ac.uk" class="auth_mail" title="E-mail the corresponding author ; Martin Taká ; čc ; Takac.MT@gmail.com" class="auth_mail" title="E-mail the corresponding author
- 关键词：Matrix completion ; Robust optimization ; Collaborative filtering ; Coordinate descent ; Large-scale optimization
- 刊名：European Journal of Operational Research
- 出版年：2017
- 出版时间：1 January 2017
- 年：2017
- 卷：256
- 期：1
- 页码：35-43
- 全文大小：1405 K
文摘
- •
Matrix completion under interval uncertainty can be encoded using box constraints.
- •
E.g. in collaborative filtering, a rating of 2 can be replaced by an interval 1–3.
- •
Present-best results in terms of statistical performance are provided for a benchmark.
- •
Convergence of randomized coordinate-descent methods to stationary points is proven.