A Regularized Nonnegative Third Order Tensor decomposition Using a Primal-Dual Projected Gradient Algorithm: Application to 3D Fluorescence Spectroscopy.
Karima El QateMohammed El RhabiAbdelilah HakimEric MoreauNadège Thirion-MoreauPublished in: ICSM (2018)
Keyphrases
- primal dual
- linear programming
- semidefinite programming
- convergence rate
- objective function
- learning algorithm
- expectation maximization
- computational complexity
- optimal solution
- k means
- nonnegative matrix factorization
- dynamic programming
- simplex method
- probabilistic model
- active learning
- lower bound
- clustering algorithm