A generic approach to nonparametric function estimation with mixed data

Thomas Nagler · 2018
PDF · DOI: 10.1016/j.spl.2018.02.040 · arXiv:1704.07457

The paper contains a nice literature overview.

Abstract

Most nonparametric function estimators can only handle continuous data. We show that making discrete variables continuous by adding noise is justified under suitable conditions on the noise distribution. This principle is widely applicable, including density and regression function estimation.

Reference

Thomas Nagler “A generic approach to nonparametric function estimation with mixed data” (2018) DOI: 10.1016/j.spl.2018.02.040 arXiv:1704.07457

@Article{nagler2018,
  title = {A generic approach to nonparametric function estimation with mixed data},
  abstract = {Most nonparametric function estimators can only handle continuous data. We show that making discrete variables continuous by adding noise is justified under suitable conditions on the noise distribution. This principle is widely applicable, including density and regression function estimation.},
  volume = {137},
  issn = {0167-7152},
  doi = {10.1016/j.spl.2018.02.040},
  arxiv = {1704.07457},
  journal = {Statistics \& Probability Letters},
  publisher = {Elsevier BV},
  author = {Nagler, Thomas},
  year = {2018},
  month = {jun},
  pages = {326–330}
}