Sufficient Component Analysis for Supervised Dimension Reduction
By Makoto Yamada, Gang Niu, Jun Takagi, Masashi Sugiyama. 2011
The purpose of sufficient dimension reduction (SDR) is to find the low-dimensional subspace of input features that is sufficient for predicting output values. In this paper, we propose a novel distribution-free SDR method called sufficient component analysis (SCA), which is computationally more efficient than existing methods. In our method, a solution is computed by iteratively performing dependence estimation and maximization: Dependence estimation is analytically carried out by recently-proposed least-squares mutual information (LSMI), and dependence maximization is also analytically carried out by utilizing the Epanechnikov kernel. Through large-scale experiments on real-world image classification and audio tagging problems, the proposed method is shown to compare favorably with existing dimension reduction approaches.
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
- Praneeth Vepakomma
- Ahmed M. Elgammal
- Chetan Tonde
- Thierry Bouwmans
- Vivek Sharma
- Olivier Barnich
- Abhishek Singh
- Ramesh Raskar
Citations in the corpus
- Supervised Dimensionality Reduction via Distance Correlation Maximization (2016, methodology): “…24.33 23.90 23.62 LSDR [Suzuki and Sugiyama, 2013] 30.36 28.16 27.39 27.24 27.18 gKDR [Fukumizu and Leng, 2014] 29.72 27.62 27.29 26.91 26.81 SCA [Yamada et al., 2011] 28.53 27.31 26.60 26.32 26.30 LAD [Cook and Forzani, 2009] 30.42 30.39 30.20 30.04 29.99 SAVE [Shao et al., 2009] 31.93 31.27…”