A -saturated model has the one-point extension property for partial elementary maps of size below : transport the type of the new element through the map and realize it by saturation. This is -homogeneity; back-and-forth gives the corresponding automorphisms when cardinalities permit.
If an aleph-zero-homogeneous model realizes every finite type over the empty set, it is aleph-zero-saturated. Realize the joint empty-set type of the parameters and a desired element, then use homogeneity to move the realized parameter tuple to the given one.
A model of cardinality is universal when every model of the same complete theory having cardinality less than admits an elementary embedding into .
A model of cardinality is homogeneous when every partial elementary map between subsets of size less than extends to an automorphism of .
A universal homogeneous model realizes every type over a small parameter set. Embed a small model containing a realization into the universal model, then use homogeneity to move the embedded parameters back to the original ones.
Let be uncountable and saturated, and let a formula define an equivalence relation. If every small elementary submodel of meets every class, then there are only finitely many classes. Otherwise a small elementary submodel yields the finitely satisfiable type , contradicting saturation.
Articles by others on the same topic
A **saturated model** is a statistical model that is fully specified to account for all possible variability in the data. In essence, it includes as many parameters as there are data points, meaning that it can fit the data perfectly. Thus, every possible outcome in the dataset is accounted for by a unique parameter within the model. Here are some key points about saturated models: 1. **Overparameterization**: Saturated models typically have a high number of parameters, making them overparameterized.