A Functional Equality with Application to Nonparametric Estimation
碩士 === 國立中正大學 === 統計科學所 === 95 === Traditional kernel estimation is applicable for obtaining nonparametric estimate of a probability density function when sample size is large. This work introduces a new equality which is applicable for improving kernel estimation to achieve improved mean square err...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | en_US |
| Published: |
2006
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/52844709702724051634 |
| Summary: | 碩士 === 國立中正大學 === 統計科學所 === 95 === Traditional kernel estimation is applicable for obtaining nonparametric estimate of a probability density function when sample size is large. This work introduces a new equality which is applicable for improving kernel estimation to achieve improved mean square error when sample size is small. We use this new method to estimate density function, hazard rate function
and regression function. The choice of bandwidth for the new kernel method on estimating the above functions is investigated, and empirical comparison is provided.
|
|---|