Department of Electrical Engineering (ISY)

Improving Random Forests by correlation-enhancing projections and sample-based sparse discriminant selection

Marcus Wallenberg, Per-Erik Forssén
CRV16, Victoria, BC
13th Conference on Computer and Robot Vision (CRV16)
June 2016

Abstract

Random Forests (RF) is a learning technique with very low run-time complexity. It has found a niche application in situations where input data is low-dimensional and computational performance is paramount. We wish to make RFs more useful for high dimensional problems, and to this end, we propose two extensions to RFs: Firstly, a feature selection mechanism called correlation-enhancing projections, and secondly sparse discriminant selection schemes for better accuracy and faster training. We evaluate the proposed extensions by performing age and gender estimation on the MORPH-II dataset, and demonstrate near-equal or improved estimation performance when using these extensions despite a seventy-fold reduction in the number of data dimensions.

Full Paper

Portable document format file PDF ()
This paper is also available from IEEE Xplore.

---

Bibtex entry

@InProceedings{wallenberg16,
  author = 	 {Marcus Wallenberg and Per-Erik Forss\'en},
  title = 	 {Improving Random Forests by correlation-enhancing projections and sample-based sparse discriminant selection},
  booktitle = {13th Conference on Computer and Robot Vision ({CRV}16)},
  year = 	 {2016},
  month = 	 {June},
  address = 	 {Victoria, BC},
  publisher = {{IEEE}}
}