Resistive Grid Kernel Estimator (RGKE)
Abstract
The ability to estimate a probability density function from random data has applications in discriminant analysis and pattern recognition problems. A resistive grid kernel estimator (RGKE) is described that is suitable for hardware implementation. The one-dimensional linear RGKE is compared to a kernel estimate using Gaussian kernels, and simulations are presented using both continuous and quantized data. The nonlinear form of the RGKE is shown to have desirable properties, such as the ability to detect discontinuities in the density function.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1992
- Accession Number
- ADA253520
Entities
People
- Carey E. Priebe
- George Rogers
- Jeffrey Solka
- Wendy Poston
Organizations
- Naval Surface Warfare Center