Last updated 2 years ago
Expected gain (MEG) = arg max(α N(x;μT,σT)−(1−α) N(x;μNT,σNT))arg\:max(\alpha\:N(x; \mu_T,\sigma_T)-(1-\alpha)\:N(x;\mu_{NT},\sigma_{NT}))argmax(αN(x;μT,σT)−(1−α)N(x;μNT,σNT))
Link to paper
