Feature Screening for Nonparametric and Semiparametric Models with Ultrahigh-Dimensional Covariates
- 103 Downloads
This paper considers the feature screening and variable selection for ultrahigh dimensional covariates. The new feature screening procedure base on conditional expectation which is used to differentiate whether an explanatory variable contributes to a response variable or not, without requiring a specific parametric form of the underlying data model. The authors estimate the marginal conditional expectation by kernel regression estimator. The proposed method is showed to have sure screen property. The authors propose an iterative kernel estimator algorithm to reduce the ultrahigh dimensionality to an appropriate scale. Simulation results and real data analysis demonstrate the proposed method works well and performs better than competing methods.
KeywordsConditional expectation dimensionality reduction nonparametric and semiparametric models ultrahigh dimension variable screening
Unable to display preview. Download preview PDF.
- Fan J, Samworth R, and Wu Y, Ultra-dimensional variable selection via independent learning: Beyond the linear model, Journal of Machine Learning Research, 2009, 10: 1829–1853.Google Scholar
- Chiang A, Beck J, Yen H, et al, Homozygosity mapping with SNP arrays identifies TRIM32, an E3 ubiquitin ligase, as a bardetiedl syndrome gene (BBS11), Proceedings of the National Academy of Sciences, 2006, 103: 6287–6292.Google Scholar