MathDB3 problems
- 0 votes0 replies0 views
Longer MCMC chains improve estimation after simulation failures
Longer-chain improvement conjecture. Running these chains longer is likely to result in improved estimation.
- 0 votes0 replies0 views
Universal full asymptotic optimality conjecture for Mallows model averaging
Let the full asymptotic optimality property (full AOP) denote achieving the best performance of a convex combination of all the subset models, uniformly over the whole parameter re…
- 0 votes0 replies0 views
The OGA-FAMMA logarithmic-factor conjecture for model averaging
OGA-FAMMA conjecture. When the variables are preordered by OGA, this minimax optimal rate should be achievable by FAMMA with the factor replaced by a factor directly proportion…