An algorithm for the recovery of both target joint beliefs and full belief from Bayesian networks.
Mark BloemekePublished in: ACM Southeast Regional Conference (1998)
Keyphrases
- learning algorithm
- experimental evaluation
- bayesian networks
- objective function
- np hard
- k means
- preprocessing
- computational cost
- times faster
- theoretical analysis
- significant improvement
- search space
- computational complexity
- optimal solution
- similarity measure
- recovery algorithm
- improved algorithm
- recognition algorithm
- detection algorithm
- worst case
- dynamic programming
- cost function
- simulated annealing
- particle swarm optimization
- em algorithm
- high accuracy
- segmentation algorithm
- matching algorithm
- decision making
- convergence rate
- conditional probabilities
- neural network
- target tracking