ABSTRACT
This chapter focuses on the two important components of a metamodeling approach: the experimental design method and the predictive model. It reviews some metamodeling approaches that have been proposed and studied in the literature for the valuation of variable annuities. In simulation metamodeling, a metamodel refers to a model of a simulation model, which is usually very complex and computationally intensive. A metamodel is constructed by running a small number of expensive simulations and used in place of the simulation model for further analysis. Metamodeling approaches are similar to inforce compression techniques introduced in that both reduce runtime through cutting down the number of policies that are valued by the valuation model. The chapter discusses metamodeling approaches for speeding up variable annuity valuation. Metamodeling stems from computer simulation modeling. A metamodel is simple and computationally efficient, a metamodeling approach can reduce the runtime of the simulation model significantly.
