Imprecise probability models for learning multinomial distributions from data. Applications to learning credal networks

Andrés R. Masegosa, Serafín Moral

Abstract: Highlights•We examine imprecise models for estimating multinomial probabilities.•The selection of the equivalent sample size in a Dirichlet density is not simple.•An imprecise model can be considered by being imprecise in the equivalent sample size.•Imprecise sample size is useful to learn credal networks with imprecise structure.