Interactive data analysis with statistical guarantees
Abstract
Major Goals: This project considered the problem of inference (with frequentist guarantees) when statisticalmodels are not specified before data collection. We termed this type of data analysis "inferactive data analysis".Two of the main goals of this project involved working out a framework for inference, the second beingdevelopment of software usable in this context. In terms of the framework, questions include: under what kind ofexploratory regimes is it possible to do inference; what are the appropriate corrections to the likelihood? Forcomputations we considered specific randomization mechanisms that allowed tractable inference algorithms.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 05, 2019
- Accession Number
- AD1083900
Entities
People
- Jonathan E. Taylor
Organizations
- Stanford University