Mathematical and Statistical Methods for Actuarial Sciences and Finance

Lee-Carter error matrix simulation 117

Fig. 2.κtresulting from different experimental conditions: the age group (on the rows) and the
different percentiles (on the columns) effect

the experiment in the four rows and the successive increment in the percentage of outer
values which have been transformed in the 6 columns. We can notice the differentκt
behaviour in the four rows as more age groups and percentiles are considered.
For better interpretation of these results, we have plotted a synthetic view of the
resulting average of the 1000κtunder the 24 conditions (see Fig. 3) and compared
them with the series derived by the traditional LC estimation.

Fig. 3.A comparison between the 24 averagedκt(in red) and the original one (in black)