A Precise High-Dimensional Asymptotic Theory for Boosting and Min-L1-Norm Interpolated Classifiers.
Tengyuan LiangPragya SurPublished in: CoRR (2020)
Keyphrases
- high dimensional
- ensemble learning
- boosting algorithms
- improving classification accuracy
- multiclass classification
- ensemble classifier
- weak classifiers
- training samples
- low dimensional
- feature selection
- boosting framework
- weak learners
- linear classifiers
- similarity search
- dimensionality reduction
- training data
- learning algorithm
- laplace transform
- randomized trees
- decision trees
- decision stumps
- multi class
- variable selection
- high dimensional data
- machine learning algorithms
- feature set
- large deviations
- adaboost algorithm
- training set
- data points
- support vector machine
- feature space
- binary classification problems
- svm classifier
- classification trees
- class labels
- principal component analysis
- boosted classifiers
- nearest neighbor