Importance Sampling for Tail Probabilities
Abstract
Our interest is in the accurate and efficient evaluation of very small probabilities. Thus, we seek methods with bounded coefficient of variation as the probability decreases and we assess their performance. The univariate case is explored for a heuristic for construction of sampling density families. The multivariate case provides interesting connections with Mills' ratio.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 06, 1991
- Accession Number
- ADA232412
Entities
People
- Satish Iyengar
Organizations
- Stanford University