A Data Based Algorithm for Choosing the Window Width when Estimating the Integral of f2(x).
Abstract
A review of different density estimation approaches is given. A new data-based algorithm for choosing the window width when estimating gamma using methods of kernel density estimation is described. The results of a Monte Carlo study comparing the data-based window widths produced by the new algorithm with those produced by an existing method are also presented. Finally, an estimator which is stable over a broad set of sampling situations is recommended. Keywords: Kernel density estimation; Window width; Data-based algorithm.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1985
- Accession Number
- ADA152093
Entities
People
- S. J. Sheather
- T. P. Hettmansperger
Organizations
- Pennsylvania State University