A Scalable MIP-based Method for Learning Optimal Multivariate Decision Trees.
Haoran ZhuPavankumar MuraliDzung T. PhanLam M. NguyenJayant KalagnanamPublished in: NeurIPS (2020)
Keyphrases
- detection method
- dynamic programming
- synthetic data
- classification accuracy
- neural network
- learning mechanism
- learning scheme
- support vector machine svm
- unsupervised learning
- high accuracy
- computational cost
- high precision
- supervised learning
- objective function
- experimental evaluation
- cost function
- significant improvement
- active learning
- prior knowledge
- pairwise
- optimal path
- data sets
- training process
- learning tasks
- exhaustive search
- multivariate decision trees
- segmentation method
- em algorithm
- online learning
- support vector machine
- evolutionary algorithm
- learning process
- optimal solution
- reinforcement learning
- feature extraction