Maximum likelihood method for bandwidth selection in kernel conditional density estimate
- 51 Downloads
This paper discusses the kernel estimator of conditional density. A significant problem of kernel smoothing is bandwidth selection. The problem consists in the fact that optimal bandwidth depends on the unknown conditional and marginal density. This is the reason why some data-driven method needs to be applied. In this paper, we suggest a method for bandwidth selection based on a classical maximum likelihood approach. We consider a slight modification of the original method—the maximum likelihood method with one observation being left out. Applied to two types of conditional density estimators—to the Nadaraya–Watson and local linear estimator, the proposed method is compared with other known methods in a simulation study. Our aim is to compare the methods from different points of view, concentrating on the accuracy of the estimated bandwidths, on the final model quality measure, and on the computational time.
KeywordsKernel smoothing Conditional density Methods for bandwidth selection Leave-one-out maximum likelihood method
The research was supported by Grant Agency of Masaryk University grant No. MUNI/A/1503/2018, and, moreover, by the project of specific university research at Brno University of Technology FAST-S-19-5878.
- Chen X, Linton O, Robinson P (2001) The estimation of conditional densities. The suntory centre discussion paper no. Tech. rep., EM/01/415, LSE, LondonGoogle Scholar
- Hansen BE (2004) Nonparametric conditional density estimation. Unpublished manuscript. https://www.ssc.wisc.edu/~bhansen/papers/ncde.pdf
- Holmes MP, Gray AG, Isbell CL (2012) Fast nonparametric conditional density estimation. ArXiv e-prints arXiv:1206.5278
- Koláček J, Horová I (2012) Iterative bandwidth method for kernel regression. J Stat Adv Theory Appl 8:93–101Google Scholar
- Konečná K, Horová I (2014) Conditional density estimations. International Society for the Advancement of Science and Technology (ISAST), Athens, pp 15–31Google Scholar
- Konečná K, Horová I (2017) Methods for bandwidth detection in kernel conditional density estimations. In: European Young Statisticians Meeting, pp 143–148Google Scholar
- Schuster EF, Gregory GG (1981) On the nonconsistency of maximum likelihood nonparametric density estimators. In: Computer Science and Statistics: Proceedings of the 13th Symposium on the interface. Springer, pp 295–298Google Scholar
- van der Walt CM, Barnard E (2017) Variable kernel density estimation in high-dimensional feature spaces. In: Thirty-first AAAI conference on artificial intelligenceGoogle Scholar