Relative error in approximating the log-likelihood via importance sampling.
2021
academicJournal
Zugriff:
The relative error is computed by dividing the absolute error by the absolute value of the true log-likelihood. We note that for the forward-pool sampling, the relative absolute error of the log-likelihood depends almost exclusively on the the size of the initial pool K for L ≥ 10. Similarly, the approximation accuracy of the Hamming k -neighborhood sampling is primarily determined by the extent k of the considered neighborhood. (PDF)
Titel: |
Relative error in approximating the log-likelihood via importance sampling.
|
---|---|
Autor/in / Beteiligte Person: | Susana Posada-Céspedes (11402519) ; Gert Van Zyl (11402522) ; Hesam Montazeri (451798) ; Jack Kuipers (1390278) ; Soo-Yon Rhee (42412) ; Roger Kouyos (178908) ; Huldrych F. Günthard (9623497) ; Niko Beerenwinkel (54419) |
Link: | |
Veröffentlichung: | 2021 |
Medientyp: | academicJournal |
DOI: | 10.1371/journal.pcbi.1008363.s018 |
Schlagwort: |
|
Sonstiges: |
|