EMPRR: a high-dimensional EM-based peicewise regression algorithm.
Manimozhiyan ArumugamStephen D. ScottPublished in: ICMLA (2004)
Keyphrases
- regression algorithm
- high dimensional
- semi supervised
- expectation maximization
- regression methods
- piecewise linear
- low dimensional
- probabilistic model
- high dimensional data
- learning algorithm
- dimensionality reduction
- unsupervised learning
- nearest neighbor
- variable selection
- data points
- em algorithm
- input space
- generative model
- high dimensionality
- neural network
- dimension reduction
- maximum likelihood
- unlabeled data
- feature selection algorithms
- co training
- feature space