Non-removal strategy for outliers in predictive models: The PAELLA algorithm case.
Manuel Castejón LimasHéctor Alaiz-MoretónLaura Fernández-RoblesJavier Alfonso-CendónCamino Fernández LlamasLidia Sánchez-GonzálezHilde PérezPublished in: Log. J. IGPL (2020)
Keyphrases
- improved algorithm
- computational cost
- probabilistic model
- objective function
- optimization algorithm
- learning algorithm
- cost function
- detection algorithm
- np hard
- experimental evaluation
- times faster
- theoretical analysis
- expectation maximization
- bayesian framework
- computational complexity
- recognition algorithm
- robust statistical
- linear programming
- model selection
- particle swarm optimization
- high accuracy
- worst case
- significant improvement
- preprocessing
- selection algorithm
- image segmentation
- predictive model
- learned models
- decision trees