Large-Scale Inference: Empirical Bayes Methods for Estimation, Testing, and Prediction (Institute of Mathematical Statistics Monographs)

Large-Scale Inference: Empirical Bayes Methods for Estimation, Testing, and Prediction (Institute of Mathematical Statistics Monographs)

Bradley Efron


we are living in a brand new age for statistical inference, the place smooth medical expertise corresponding to microarrays and fMRI machines mostly produce hundreds of thousands and infrequently hundreds of thousands of parallel facts units, each one with its personal estimation or trying out challenge. Doing hundreds of thousands of difficulties instantly is greater than repeated software of classical tools. Taking an empirical Bayes strategy, Bradley Efron, inventor of the bootstrap, indicates how details accrues throughout difficulties in a fashion that mixes Bayesian and frequentist principles. Estimation, trying out, and prediction combination during this framework, generating possibilities for brand spanking new methodologies of elevated energy. New problems additionally come up, simply resulting in incorrect inferences. This ebook takes a cautious examine either the promise and pitfalls of large-scale statistical inference, with specific recognition to fake discovery charges, the main profitable of the hot statistical strategies. Emphasis is at the inferential rules underlying technical advancements, illustrated utilizing quite a few actual examples.

Show sample text content

Download sample