Newton-based optimization for Kullback–Leibler nonnegative tensor factorizations
- 30 April 2015
- journal article
- research article
- Published by Taylor & Francis in Optimization Methods and Software
- Vol. 30 (5) , 1002-1029
Abstract
Tensor factorizations with nonnegativity constraints have found application in analysing data from cyber traffic, social networks, and other areas. We consider application data best described as being generated by a Poisson process (e.g. count data), which leads to sparse tensors that can be modelled by sparse factor matrices. In this paper, we investigate efficient techniques for computing an appropriate canonical polyadic tensor factorization based on the Kullback–Leibler divergence function. We propose novel subproblem solvers within the standard alternating block variable approach. Our new methods exploit structure and reformulate the optimization problem as small independent subproblems. We employ bound-constrained Newton and quasi-Newton methods. We compare our algorithms against other codes, demonstrating superior speed for high accuracy results and the ability to quickly find sparse solutions.Keywords
All Related Versions
This publication has 37 references indexed in Scilit:
- On Tensors, Sparsity, and Nonnegative FactorizationsSIAM Journal on Matrix Analysis and Applications, 2012
- Temporal Link Prediction Using Matrix and Tensor FactorizationsACM Transactions on Knowledge Discovery From Data, 2011
- A scalable optimization approach for fitting canonical tensor decompositionsJournal of Chemometrics, 2011
- Fast Local Algorithms for Large Scale Nonnegative Matrix and Tensor FactorizationsIEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences, 2009
- A fast non-negativity-constrained least squares algorithmJournal of Chemometrics, 1997
- A Limited Memory Algorithm for Bound Constrained OptimizationSIAM Journal on Scientific Computing, 1995
- Global Convergence of a Class of Trust Region Algorithms for Optimization with Simple BoundsSIAM Journal on Numerical Analysis, 1988
- Projected Newton Methods for Optimization Problems with Simple ConstraintsSIAM Journal on Control and Optimization, 1982
- On the Goldstein-Levitin-Polyak gradient projection methodIEEE Transactions on Automatic Control, 1976
- Analysis of Individual Differences in Multidimensional Scaling Via an N-way Generalization of “Eckart-Young” DecompositionPsychometrika, 1970