Robust normal reference bandwidth for kernel density estimation |
| |
| Authors: | Jin Zhang Xueren Wang |
| |
| Affiliation: | School of Mathematics and Statistics, Yunnan University, Kunming, Yunnan, 650091, China |
| |
| Abstract: | Bandwidth selection is the main problem of kernel density estimation, the most popular method of density estimation. The classical normal reference bandwidth usually oversmoothes the density estimate. The existing hi-tech bandwidths have computational problems (even may not exist) and are not robust against outliers in the sample. A highly robust normal reference bandwidth is proposed, which adapts to different types of densities. |
| |
| Keywords: | biased cross-validation least squares cross-validation mean integrated squared error outliers |
|
|
