I like the approach, but the random N17 exploration is not ideal. The characterized sequences don’t form many inhibitory RNA structures and therefore the ML doesn’t acquire that *known* sequence-activity pattern/interaction.
-
-
-
If an ML model can’t recapitulate a known sequence-activity interaction, then its accuracy metrics are meaningless. The metric is always conditioned on what’s in the dataset ....
- Još 4 druga odgovora
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.