diblc: Dynamic Improvement Bayesian Lee-Carter

In vsart/dynamicimprovement: Dynamic Improvement Mortality Modelling

Description

Performs estimation of a Bayesian Lee-Carter model for modelling mortality date by extending the research done by Pedroza (2006) where Dynamic Linear Models were used for performing joint inference on the model of Lee & Carter (1992). This routine employes empirical estimates for the 'alpha' parameter and introduces a dynamic to the improvement terms, which have evolutional variances modelled through discount factors, as suggested in West & Harrison (1996). This is the result of ongoing research, and this function may change with time to reflect the latest model considered the best approach by the authors.

Usage

diblc(Y, I = 3000, B = 500, df_b = 0.9)

Arguments

Y	Numerical matrix with age information on rows and time information on columns. Must be log-mortality with valid and finite values.
I	Number of iterations for the Gibbs sampler.
B	Number of iterations to be discarded.
df_b	Discount factor for improvement parameters. Defaults to 0.9.

Value

A 'diblc' object.

vsart/dynamicimprovement documentation built on May 26, 2019, 5:35 a.m.