A generalization of the Remez algorithm to a class of linear spline approximation problems with constraints on spline parameters.
Nadezda SukhorukovaPublished in: Optim. Methods Softw. (2008)
Keyphrases
- expectation maximization
- np hard
- learning algorithm
- convex quadratic programming
- benchmark problems
- cost function
- dynamic programming
- parameter estimation
- matching algorithm
- detection algorithm
- boolean expressions
- constrained problems
- path consistency
- continuous variables
- input data
- linear programming
- least squares
- image segmentation
- linear least squares
- computational complexity
- linear systems
- constrained optimization
- parameter settings
- worst case
- combinatorial optimization
- closed form
- maximum likelihood
- em algorithm
- multi class classification
- optimal solution
- objective function
- constraint sets
- highly non linear
- np complete