Nonnegative Sparse Recovery under Sum-to-One Constraint via Projected Gradient Descent
The invention is a novel approach to recovering a sparse signal that satisfies two key constraints: nonnegativity and the “sum-to-one” constraint.
Prof. SO Hing Cheung, Prof. LEUNG Chi Sing Andrew, Dr. LI Xiaopeng
- Nonnegative sparse recovery
- Sum-to-one
- Algorithm
- Gradient descent
- Projection
- Signal processing
- Algorithmic Trading
- Hyperspectral imaging
- Artificial Intelligence
- Sparse portfolio