Bootstrapping signs and permutations for regression with heavy-tailed errors: A robust resampling. (English) Zbl 0922.62027
Adler, Robert J. (ed.) et al., A practical guide to heavy tails. Statistical techniques and applications. Boston: Birkhäuser. 339-358 (1998).
Summary: For multiple linear regression with heavy-tailed errors, we describe and discuss in detail two resampling methods: one based on resampling permutations of residuals from the least squares fit, the other, to some extent a complementary one, exploiting random sign flips. Both succeed, without moment assumptions, in recovering conditional distributions of the least squares estimators. We outline consequences of the results for effective statistical inference for regression models with outsized errors.
For the entire collection see [Zbl 0901.00010].
For the entire collection see [Zbl 0901.00010].
