Graduating the age-specific fertility pattern using Support Vector Machines

A topic of interest in demographic literature is the graduation of the age-specific fertility pattern. A classical graduation technique extensively used by demographers is to fit parametric models that accurately reproduce it. Standard non parametric statistical methodology, as kernels and splines, might alternately be used for this graduation purpose. Support Vector Machines (SVM) is an innovative non parametric methodology that could also be used for fertility graduation purposes. This paper evaluates SVM techniques as tools for graduating fertility rates. To that end, we apply these techniques to empirical age-specific fertility rates from a variety of populations and time periods. Additionally, for comparison reasons we also fit parametric models and kernels to these empirical data sets.